Computer Vision and Pattern Recognition (cs.CV)

Abstract: Lip-based biometric authentication (LBBA) is an authentication method based on a person's lip movements during speech in the form of video data captured by a camera sensor. LBBA can utilize both physical and behavioral characteristics of lip movements without requiring any additional sensory equipment apart from an RGB camera. State-of-the-art (SOTA) approaches use one-shot learning to train deep siamese neural networks which produce an embedding vector out of these features. Embeddings are further used to compute the similarity between an enrolled user and a user being authenticated. A flaw of these approaches is that they model behavioral features as style-of-speech without relation to what is being said. This makes the system vulnerable to video replay attacks of the client speaking any phrase. To solve this problem we propose a one-shot approach which models behavioral features to discriminate against what is being said in addition to style-of-speech. We achieve this by customizing the GRID dataset to obtain required triplets and training a siamese neural network based on 3D convolutions and recurrent neural network layers. A custom triplet loss for batch-wise hard-negative mining is proposed. Obtained results using an open-set protocol are 3.2% FAR and 3.8% FRR on the test set of the customized GRID dataset. Additional analysis of the results was done to quantify the influence and discriminatory power of behavioral and physical features for LBBA.

Abstract: Aerial-to-ground image synthesis is an emerging and challenging problem that aims to synthesize a ground image from an aerial image. Due to the highly different layout and object representation between the aerial and ground images, existing approaches usually fail to transfer the components of the aerial scene into the ground scene. In this paper, we propose a novel framework to explore the challenges by imposing enhanced structural alignment and semantic awareness. We introduce a novel semantic-attentive feature transformation module that allows to reconstruct the complex geographic structures by aligning the aerial feature to the ground layout. Furthermore, we propose semantic-aware loss functions by leveraging a pre-trained segmentation network. The network is enforced to synthesize realistic objects across various classes by separately calculating losses for different classes and balancing them. Extensive experiments including comparisons with previous methods and ablation studies show the effectiveness of the proposed framework both qualitatively and quantitatively.

Abstract: Recent DETR-based video grounding models have made the model directly predict moment timestamps without any hand-crafted components, such as a pre-defined proposal or non-maximum suppression, by learning moment queries. However, their input-agnostic moment queries inevitably overlook an intrinsic temporal structure of a video, providing limited positional information. In this paper, we formulate an event-aware dynamic moment query to enable the model to take the input-specific content and positional information of the video into account. To this end, we present two levels of reasoning: 1) Event reasoning that captures distinctive event units constituting a given video using a slot attention mechanism; and 2) moment reasoning that fuses the moment queries with a given sentence through a gated fusion transformer layer and learns interactions between the moment queries and video-sentence representations to predict moment timestamps. Extensive experiments demonstrate the effectiveness and efficiency of the event-aware dynamic moment queries, outperforming state-of-the-art approaches on several video grounding benchmarks.

Abstract: The exponential growth of data, alongside advancements in model structures and loss functions, has necessitated the enhancement of image retrieval systems through the utilization of new models with superior feature embeddings. However, the expensive process of updating the old retrieval database by replacing embeddings poses a challenge. As a solution, backward-compatible training can be employed to avoid the necessity of updating old retrieval datasets. While previous methods achieved backward compatibility by aligning prototypes of the old model, they often overlooked the distribution of the old features, thus limiting their effectiveness when the old model's low quality leads to a weakly discriminative feature distribution. On the other hand, instance-based methods like L2 regression take into account the distribution of old features but impose strong constraints on the performance of the new model itself. In this paper, we propose MixBCT, a simple yet highly effective backward-compatible training method that serves as a unified framework for old models of varying qualities. Specifically, we summarize four constraints that are essential for ensuring backward compatibility in an ideal scenario, and we construct a single loss function to facilitate backward-compatible training. Our approach adaptively adjusts the constraint domain for new features based on the distribution of the old embeddings. We conducted extensive experiments on the large-scale face recognition datasets MS1Mv3 and IJB-C to verify the effectiveness of our method. The experimental results clearly demonstrate its superiority over previous methods. Code is available at https://github.com/yuleung/MixBCT

Abstract: Utilizing a two-stage paradigm comprising of coarse image retrieval and precise reranking, a well-established image retrieval system is formed. It has been widely accepted for long time that local feature is imperative to the subsequent stage - reranking, but this requires sizeable storage and computing capacities. We, for the first time, propose an image retrieval paradigm leveraging global feature only to enable accurate and lightweight image retrieval for both coarse retrieval and reranking, thus the name - SuperGlobal. It consists of several plug-in modules that can be easily integrated into an already trained model, for both coarse retrieval and reranking stage. This series of approaches is inspired by the investigation into Generalized Mean (GeM) Pooling. Possessing these tools, we strive to defy the notion that local feature is essential for a high-performance image retrieval paradigm. Extensive experiments demonstrate substantial improvements compared to the state of the art in standard benchmarks. Notably, on the Revisited Oxford (ROxford)+1M Hard dataset, our single-stage results improve by 8.2% absolute, while our two-stage version gain reaches 3.7% with a strong 7568X speedup. Furthermore, when the full SuperGlobal is compared with the current single-stage state-of-the-art method, we achieve roughly 17% improvement with a minimal 0.005% time overhead. Code: https://github.com/ShihaoShao-GH/SuperGlobal.

Abstract: The reconstruction of object surfaces from multi-view images or monocular video is a fundamental issue in computer vision. However, much of the recent research concentrates on reconstructing geometry through implicit or explicit methods. In this paper, we shift our focus towards reconstructing mesh in conjunction with color. We remove the view-dependent color from neural volume rendering while retaining volume rendering performance through a relighting network. Mesh is extracted from the signed distance function (SDF) network for the surface, and color for each surface vertex is drawn from the global color network. To evaluate our approach, we conceived a in hand object scanning task featuring numerous occlusions and dramatic shifts in lighting conditions. We've gathered several videos for this task, and the results surpass those of any existing methods capable of reconstructing mesh alongside color. Additionally, our method's performance was assessed using public datasets, including DTU, BlendedMVS, and OmniObject3D. The results indicated that our method performs well across all these datasets. Project page: https://colmar-zlicheng.github.io/color_neus.

Abstract: 3D object reconstruction and multilevel segmentation are fundamental to computer vision research. Existing algorithms usually perform 3D scene reconstruction and target objects segmentation independently, and the performance is not fully guaranteed due to the challenge of the 3D segmentation. Here we propose an open-source one stop 3D target reconstruction and multilevel segmentation framework (OSTRA), which performs segmentation on 2D images, tracks multiple instances with segmentation labels in the image sequence, and then reconstructs labelled 3D objects or multiple parts with Multi-View Stereo (MVS) or RGBD-based 3D reconstruction methods. We extend object tracking and 3D reconstruction algorithms to support continuous segmentation labels to leverage the advances in the 2D image segmentation, especially the Segment-Anything Model (SAM) which uses the pretrained neural network without additional training for new scenes, for 3D object segmentation. OSTRA supports most popular 3D object models including point cloud, mesh and voxel, and achieves high performance for semantic segmentation, instance segmentation and part segmentation on several 3D datasets. It even surpasses the manual segmentation in scenes with complex structures and occlusions. Our method opens up a new avenue for reconstructing 3D targets embedded with rich multi-scale segmentation information in complex scenes. OSTRA is available from https://github.com/ganlab/OSTRA.

Abstract: Nearest neighbor (NN) sampling provides more semantic variations than pre-defined transformations for self-supervised learning (SSL) based image recognition problems. However, its performance is restricted by the quality of the support set, which holds positive samples for the contrastive loss. In this work, we show that the quality of the support set plays a crucial role in any nearest neighbor based method for SSL. We then provide a refined baseline (pNNCLR) to the nearest neighbor based SSL approach (NNCLR). To this end, we introduce pseudo nearest neighbors (pNN) to control the quality of the support set, wherein, rather than sampling the nearest neighbors, we sample in the vicinity of hard nearest neighbors by varying the magnitude of the resultant vector and employing a stochastic sampling strategy to improve the performance. Additionally, to stabilize the effects of uncertainty in NN-based learning, we employ a smooth-weight-update approach for training the proposed network. Evaluation of the proposed method on multiple public image recognition and medical image recognition datasets shows that it performs up to 8 percent better than the baseline nearest neighbor method, and is comparable to other previously proposed SSL methods.

Abstract: Accurately detecting objects in the environment is a key challenge for autonomous vehicles. However, obtaining annotated data for detection is expensive and time-consuming. We introduce PatchContrast, a novel self-supervised point cloud pre-training framework for 3D object detection. We propose to utilize two levels of abstraction to learn discriminative representation from unlabeled data: proposal-level and patch-level. The proposal-level aims at localizing objects in relation to their surroundings, whereas the patch-level adds information about the internal connections between the object's components, hence distinguishing between different objects based on their individual components. We demonstrate how these levels can be integrated into self-supervised pre-training for various backbones to enhance the downstream 3D detection task. We show that our method outperforms existing state-of-the-art models on three commonly-used 3D detection datasets.

Abstract: Multi-stage architectures have exhibited efficacy in image dehazing, which usually decomposes a challenging task into multiple more tractable sub-tasks and progressively estimates latent hazy-free images. Despite the remarkable progress, existing methods still suffer from the following shortcomings: (1) limited exploration of frequency domain information; (2) insufficient information interaction; (3) severe feature redundancy. To remedy these issues, we propose a novel Mutual Information-driven Triple interaction Network (MITNet) based on spatial-frequency dual domain information and two-stage architecture. To be specific, the first stage, named amplitude-guided haze removal, aims to recover the amplitude spectrum of the hazy images for haze removal. And the second stage, named phase-guided structure refined, devotes to learning the transformation and refinement of the phase spectrum. To facilitate the information exchange between two stages, an Adaptive Triple Interaction Module (ATIM) is developed to simultaneously aggregate cross-domain, cross-scale, and cross-stage features, where the fused features are further used to generate content-adaptive dynamic filters so that applying them to enhance global context representation. In addition, we impose the mutual information minimization constraint on paired scale encoder and decoder features from both stages. Such an operation can effectively reduce information redundancy and enhance cross-stage feature complementarity. Extensive experiments on multiple public datasets exhibit that our MITNet performs superior performance with lower model complexity.The code and models are available at https://github.com/it-hao/MITNet.

Abstract: Adversarial camouflage has garnered attention for its ability to attack object detectors from any viewpoint by covering the entire object's surface. However, universality and robustness in existing methods often fall short as the transferability aspect is often overlooked, thus restricting their application only to a specific target with limited performance. To address these challenges, we present Adversarial Camouflage for Transferable and Intensive Vehicle Evasion (ACTIVE), a state-of-the-art physical camouflage attack framework designed to generate universal and robust adversarial camouflage capable of concealing any 3D vehicle from detectors. Our framework incorporates innovative techniques to enhance universality and robustness: a refined texture rendering that enables common texture application to different vehicles without being constrained to a specific texture map, a novel stealth loss that renders the vehicle undetectable, and a smooth and camouflage loss to enhance the naturalness of the adversarial camouflage. Our extensive experiments on 15 different models show that ACTIVE consistently outperforms existing works on various public detectors, including the latest YOLOv7. Notably, our universality evaluations reveal promising transferability to other vehicle classes, tasks (segmentation models), and the real world, not just other vehicles.

Abstract: Hyperspectral imagery contains abundant spectral information beyond the visible RGB bands, providing rich discriminative details about objects in a scene. Leveraging such data has the potential to enhance visual tracking performance. While prior hyperspectral trackers employ CNN or hybrid CNN-Transformer architectures, we propose a novel approach HPFormer on Transformers to capitalize on their powerful representation learning capabilities. The core of HPFormer is a Hyperspectral Hybrid Attention (HHA) module which unifies feature extraction and fusion within one component through token interactions. Additionally, a Transform Band Module (TBM) is introduced to selectively aggregate spatial details and spectral signatures from the full hyperspectral input for injecting informative target representations. Extensive experiments demonstrate state-of-the-art performance of HPFormer on benchmark NIR and VIS tracking datasets. Our work provides new insights into harnessing the strengths of transformers and hyperspectral fusion to advance robust object tracking.

Abstract: This paper proposes a novel unsupervised domain adaption (UDA) method based on contrastive bi-projector (CBP), which can improve the existing UDA methods. It is called CBPUDA here, which effectively promotes the feature extractors (FEs) to reduce the generation of ambiguous features for classification and domain adaption. The CBP differs from traditional bi-classifier-based methods at that these two classifiers are replaced with two projectors of performing a mapping from the input feature to two distinct features. These two projectors and the FEs in the CBPUDA can be trained adversarially to obtain more refined decision boundaries so that it can possess powerful classification performance. Two properties of the proposed loss function are analyzed here. The first property is to derive an upper bound of joint prediction entropy, which is used to form the proposed loss function, contrastive discrepancy (CD) loss. The CD loss takes the advantages of the contrastive learning and the bi-classifier. The second property is to analyze the gradient of the CD loss and then overcome the drawback of the CD loss. The result of the second property is utilized in the development of the gradient scaling (GS) scheme in this paper. The GS scheme can be exploited to tackle the unstable problem of the CD loss because training the CBPUDA requires using contrastive learning and adversarial learning at the same time. Therefore, using the CD loss with the GS scheme overcomes the problem mentioned above to make features more compact for intra-class and distinguishable for inter-class. Experimental results express that the CBPUDA is superior to conventional UDA methods under consideration in this paper for UDA and fine-grained UDA tasks.

Abstract: Fingerprint recognition on mobile devices is an important method for identity verification. However, real fingerprints usually contain sweat and moisture which leads to poor recognition performance. In addition, for rolling out slimmer and thinner phones, technology companies reduce the size of recognition sensors by embedding them with the power button. Therefore, the limited size of fingerprint data also increases the difficulty of recognition. Denoising the small-area wet fingerprint images to clean ones becomes crucial to improve recognition performance. In this paper, we propose an end-to-end trainable progressive guided multi-task neural network (PGT-Net). The PGT-Net includes a shared stage and specific multi-task stages, enabling the network to train binary and non-binary fingerprints sequentially. The binary information is regarded as guidance for output enhancement which is enriched with the ridge and valley details. Moreover, a novel residual scaling mechanism is introduced to stabilize the training process. Experiment results on the FW9395 and FT-lightnoised dataset provided by FocalTech shows that PGT-Net has promising performance on the wet-fingerprint denoising and significantly improves the fingerprint recognition rate (FRR). On the FT-lightnoised dataset, the FRR of fingerprint recognition can be declined from 17.75% to 4.47%. On the FW9395 dataset, the FRR of fingerprint recognition can be declined from 9.45% to 1.09%.

Abstract: Multimodal contrastive learning aims to train a general-purpose feature extractor, such as CLIP, on vast amounts of raw, unlabeled paired image-text data. This can greatly benefit various complex downstream tasks, including cross-modal image-text retrieval and image classification. Despite its promising prospect, the security issue of cross-modal pre-trained encoder has not been fully explored yet, especially when the pre-trained encoder is publicly available for commercial use. In this work, we propose AdvCLIP, the first attack framework for generating downstream-agnostic adversarial examples based on cross-modal pre-trained encoders. AdvCLIP aims to construct a universal adversarial patch for a set of natural images that can fool all the downstream tasks inheriting the victim cross-modal pre-trained encoder. To address the challenges of heterogeneity between different modalities and unknown downstream tasks, we first build a topological graph structure to capture the relevant positions between target samples and their neighbors. Then, we design a topology-deviation based generative adversarial network to generate a universal adversarial patch. By adding the patch to images, we minimize their embeddings similarity to different modality and perturb the sample distribution in the feature space, achieving unviersal non-targeted attacks. Our results demonstrate the excellent attack performance of AdvCLIP on two types of downstream tasks across eight datasets. We also tailor three popular defenses to mitigate AdvCLIP, highlighting the need for new defense mechanisms to defend cross-modal pre-trained encoders.

Abstract: Recently, Neural Radiance Field (NeRF) has shown great success in rendering novel-view images of a given scene by learning an implicit representation with only posed RGB images. NeRF and relevant neural field methods (e.g., neural surface representation) typically optimize a point-wise loss and make point-wise predictions, where one data point corresponds to one pixel. Unfortunately, this line of research failed to use the collective supervision of distant pixels, although it is known that pixels in an image or scene can provide rich structural information. To the best of our knowledge, we are the first to design a nonlocal multiplex training paradigm for NeRF and relevant neural field methods via a novel Stochastic Structural SIMilarity (S3IM) loss that processes multiple data points as a whole set instead of process multiple inputs independently. Our extensive experiments demonstrate the unreasonable effectiveness of S3IM in improving NeRF and neural surface representation for nearly free. The improvements of quality metrics can be particularly significant for those relatively difficult tasks: e.g., the test MSE loss unexpectedly drops by more than 90% for TensoRF and DVGO over eight novel view synthesis tasks; a 198% F-score gain and a 64% Chamfer $L_{1}$ distance reduction for NeuS over eight surface reconstruction tasks. Moreover, S3IM is consistently robust even with sparse inputs, corrupted images, and dynamic scenes.

Abstract: Rank-order coding, a form of temporal coding, has emerged as a promising scheme to explain the rapid ability of the mammalian brain. Owing to its speed as well as efficiency, rank-order coding is increasingly gaining interest in diverse research areas beyond neuroscience. However, much uncertainty still exists about the performance of rank-order coding under noise. Herein we show what information rates are fundamentally possible and what trade-offs are at stake. An unexpected finding in this paper is the emergence of a special class of errors that, in a regime, increase with less noise.

Abstract: The quantification of cognitive powers rests on identifying a behavioural task that depends on them. Such dependence cannot be assured, for the powers a task invokes cannot be experimentally controlled or constrained a priori, resulting in unknown vulnerability to failure of specificity and generalisability. Evaluating a compact version of Raven's Advanced Progressive Matrices (RAPM), a widely used clinical test of fluid intelligence, we show that LaMa, a self-supervised artificial neural network trained solely on the completion of partially masked images of natural environmental scenes, achieves human-level test scores a prima vista, without any task-specific inductive bias or training. Compared with cohorts of healthy and focally lesioned participants, LaMa exhibits human-like variation with item difficulty, and produces errors characteristic of right frontal lobe damage under degradation of its ability to integrate global spatial patterns. LaMa's narrow training and limited capacity -- comparable to the nervous system of the fruit fly -- suggest RAPM may be open to computationally simple solutions that need not necessarily invoke abstract reasoning.

Abstract: The increasing popularity of compact and inexpensive cameras, e.g.~dash cameras, body cameras, and cameras equipped on robots, has sparked a growing interest in detecting anomalies within dynamic scenes recorded by moving cameras. However, existing reviews primarily concentrate on Video Anomaly Detection (VAD) methods assuming static cameras. The VAD literature with moving cameras remains fragmented, lacking comprehensive reviews to date. To address this gap, we endeavor to present the first comprehensive survey on Moving Camera Video Anomaly Detection (MC-VAD). We delve into the research papers related to MC-VAD, critically assessing their limitations and highlighting associated challenges. Our exploration encompasses three application domains: security, urban transportation, and marine environments, which in turn cover six specific tasks. We compile an extensive list of 25 publicly-available datasets spanning four distinct environments: underwater, water surface, ground, and aerial. We summarize the types of anomalies these datasets correspond to or contain, and present five main categories of approaches for detecting such anomalies. Lastly, we identify future research directions and discuss novel contributions that could advance the field of MC-VAD. With this survey, we aim to offer a valuable reference for researchers and practitioners striving to develop and advance state-of-the-art MC-VAD methods.

Abstract: The morbidity of scalp diseases is minuscule compared to other diseases, but the impact on the patient's life is enormous. It is common for people to experience scalp problems that include Dandruff, Psoriasis, Tinea-Capitis, Alopecia and Atopic-Dermatitis. In accordance with WHO research, approximately 70% of adults have problems with their scalp. It has been demonstrated in descriptive research that hair quality is impaired by impaired scalp, but these impacts are reversible with early diagnosis and treatment. Deep Learning advances have demonstrated the effectiveness of CNN paired with FCN in diagnosing scalp and skin disorders. In one proposed Deep-Learning-based scalp inspection and diagnosis system, an imaging microscope and a trained model are combined with an app that classifies scalp disorders accurately with an average precision of 97.41%- 99.09%. Another research dealt with classifying the Psoriasis using the CNN with an accuracy of 82.9%. As part of another study, an ML based algorithm was also employed. It accurately classified the healthy scalp and alopecia areata with 91.4% and 88.9% accuracy with SVM and KNN algorithms. Using deep learning models to diagnose scalp related diseases has improved due to advancements i computation capabilities and computer vision, but there remains a wide horizon for further improvements.

Abstract: Ridge surfaces represent important features for the analysis of 3-dimensional (3D) datasets in diverse applications and are often derived from varying underlying data including flow fields, geological fault data, and point data, but they can also be present in the original scalar images acquired using a plethora of imaging techniques. Our work is motivated by the analysis of image data acquired using micro-computed tomography (Micro-CT) of ancient, rolled and folded thin-layer structures such as papyrus, parchment, and paper as well as silver and lead sheets. From these documents we know that they are 2-dimensional (2D) in nature. Hence, we are particularly interested in reconstructing 2D manifolds that approximate the document's structure. The image data from which we want to reconstruct the 2D manifolds are often very noisy and represent folded, densely-layered structures with many artifacts, such as ruptures or layer splitting and merging. Previous ridge-surface extraction methods fail to extract the desired 2D manifold for such challenging data. We have therefore developed a novel method to extract 2D manifolds. The proposed method uses a local fast marching scheme in combination with a separation of the region covered by fast marching into two sub-regions. The 2D manifold of interest is then extracted as the surface separating the two sub-regions. The local scheme can be applied for both automatic propagation as well as interactive analysis. We demonstrate the applicability and robustness of our method on both artificial data as well as real-world data including folded silver and papyrus sheets.

Abstract: This study attempts to segment teeth and root-canals simultaneously from CBCT images, but there are very challenging problems in this process. First, the clinical CBCT image data is very large (e.g., 672 *688 * 688), and the use of downsampling operation will lose useful information about teeth and root canals. Second, teeth and root canals are very different in morphology, and it is difficult for a simple network to identify them precisely. In addition, there are weak edges at the tooth, between tooth and root canal, which makes it very difficult to segment such weak edges. To this end, we propose a coarse-to-fine segmentation method based on inverse feature fusion transformer and uncertainty estimation to address above challenging problems. First, we use the downscaled volume data (e.g., 128 * 128 * 128) to conduct coarse segmentation and map it to the original volume to obtain the area of teeth and root canals. Then, we design a transformer with reverse feature fusion, which can bring better segmentation effect of different morphological objects by transferring deeper features to shallow features. Finally, we design an auxiliary branch to calculate and refine the difficult areas in order to improve the weak edge segmentation performance of teeth and root canals. Through the combined tooth and root canal segmentation experiment of 157 clinical high-resolution CBCT data, it is verified that the proposed method is superior to the existing tooth or root canal segmentation methods.

Abstract: Modern supervised semantic segmentation methods are usually finetuned based on the supervised or self-supervised models pre-trained on ImageNet. Recent work shows that transferring the knowledge from CLIP to semantic segmentation via prompt learning can achieve promising performance. The performance boost comes from the feature enhancement with multimodal alignment, i.e., the dot product between vision and text embeddings. However, how to improve the multimodal alignment for better transfer performance in dense tasks remains underexplored. In this work, we focus on improving the quality of vision-text alignment from two aspects of prompting design and loss function, and present an instance-conditioned prompting with contrastive learning (ICPC) framework. First, compared with the static prompt designs, we reveal that dynamic prompting conditioned on image content can more efficiently utilize the text encoder for complex dense tasks. Second, we propose an align-guided contrastive loss to refine the alignment of vision and text embeddings. We further propose lightweight multi-scale alignment for better performance. Extensive experiments on three large-scale datasets (ADE20K, COCO-Stuff10k, and ADE20K-Full) demonstrate that ICPC brings consistent improvements across diverse backbones. Taking ResNet-50 as an example, ICPC outperforms the state-of-the-art counterpart by 1.71%, 1.05%, and 1.41% mIoU on the three datasets, respectively.

Abstract: In 3D human action recognition, limited supervised data makes it challenging to fully tap into the modeling potential of powerful networks such as transformers. As a result, researchers have been actively investigating effective self-supervised pre-training strategies. In this work, we show that instead of following the prevalent pretext task to perform masked self-component reconstruction in human joints, explicit contextual motion modeling is key to the success of learning effective feature representation for 3D action recognition. Formally, we propose the Masked Motion Prediction (MAMP) framework. To be specific, the proposed MAMP takes as input the masked spatio-temporal skeleton sequence and predicts the corresponding temporal motion of the masked human joints. Considering the high temporal redundancy of the skeleton sequence, in our MAMP, the motion information also acts as an empirical semantic richness prior that guide the masking process, promoting better attention to semantically rich temporal regions. Extensive experiments on NTU-60, NTU-120, and PKU-MMD datasets show that the proposed MAMP pre-training substantially improves the performance of the adopted vanilla transformer, achieving state-of-the-art results without bells and whistles. The source code of our MAMP is available at https://github.com/maoyunyao/MAMP.

Abstract: This paper aims to tackle a novel task - Temporal Sentence Grounding in Streaming Videos (TSGSV). The goal of TSGSV is to evaluate the relevance between a video stream and a given sentence query. Unlike regular videos, streaming videos are acquired continuously from a particular source, and are always desired to be processed on-the-fly in many applications such as surveillance and live-stream analysis. Thus, TSGSV is challenging since it requires the model to infer without future frames and process long historical frames effectively, which is untouched in the early methods. To specifically address the above challenges, we propose two novel methods: (1) a TwinNet structure that enables the model to learn about upcoming events; and (2) a language-guided feature compressor that eliminates redundant visual frames and reinforces the frames that are relevant to the query. We conduct extensive experiments using ActivityNet Captions, TACoS, and MAD datasets. The results demonstrate the superiority of our proposed methods. A systematic ablation study also confirms their effectiveness.

Abstract: Key-point-based scene understanding is fundamental for autonomous driving applications. At the same time, optical flow plays an important role in many vision tasks. However, due to the implicit bias of equal attention on all points, classic data-driven optical flow estimation methods yield less satisfactory performance on key points, limiting their implementations in key-point-critical safety-relevant scenarios. To address these issues, we introduce a points-based modeling method that requires the model to learn key-point-related priors explicitly. Based on the modeling method, we present FocusFlow, a framework consisting of 1) a mix loss function combined with a classic photometric loss function and our proposed Conditional Point Control Loss (CPCL) function for diverse point-wise supervision; 2) a conditioned controlling model which substitutes the conventional feature encoder by our proposed Condition Control Encoder (CCE). CCE incorporates a Frame Feature Encoder (FFE) that extracts features from frames, a Condition Feature Encoder (CFE) that learns to control the feature extraction behavior of FFE from input masks containing information of key points, and fusion modules that transfer the controlling information between FFE and CFE. Our FocusFlow framework shows outstanding performance with up to +44.5% precision improvement on various key points such as ORB, SIFT, and even learning-based SiLK, along with exceptional scalability for most existing data-driven optical flow methods like PWC-Net, RAFT, and FlowFormer. Notably, FocusFlow yields competitive or superior performances rivaling the original models on the whole frame. The source code will be available at https://github.com/ZhonghuaYi/FocusFlow_official.

Abstract: Popular test oracles for the perception subsystem of driving automation systems identify true-positive (TP), false-positive (FP), and false-negative (FN) objects. Oracle transparency is needed for comparing test results and for safety cases. To date, there exists a common notion of TPs, FPs, and FNs in the field, but apparently no published way to comprehensively define their oracles. Therefore, this paper provides a checklist of functional aspects and implementation details that affect the oracle behavior. Besides labeling policies of the test set, we cover fields of view, occlusion handling, safety-relevant areas, matching criteria, temporal and probabilistic issues, and further aspects. Even though our checklist can hardly be formalized, it can help practitioners maximize the transparency of their oracles, which, in turn, makes statements on object perception more reliable and comparable.

Abstract: This paper presents a module, Spatial Cross-scale Convolution (SCSC), which is verified to be effective in improving both CNNs and Transformers. Nowadays, CNNs and Transformers have been successful in a variety of tasks. Especially for Transformers, increasing works achieve state-of-the-art performance in the computer vision community. Therefore, researchers start to explore the mechanism of those architectures. Large receptive fields, sparse connections, weight sharing, and dynamic weight have been considered keys to designing effective base models. However, there are still some issues to be addressed: large dense kernels and self-attention are inefficient, and large receptive fields make it hard to capture local features. Inspired by the above analyses and to solve the mentioned problems, in this paper, we design a general module taking in these design keys to enhance both CNNs and Transformers. SCSC introduces an efficient spatial cross-scale encoder and spatial embed module to capture assorted features in one layer. On the face recognition task, FaceResNet with SCSC can improve 2.7% with 68% fewer FLOPs and 79% fewer parameters. On the ImageNet classification task, Swin Transformer with SCSC can achieve even better performance with 22% fewer FLOPs, and ResNet with CSCS can improve 5.3% with similar complexity. Furthermore, a traditional network (e.g., ResNet) embedded with SCSC can match Swin Transformer's performance.

Abstract: Deep learning has achieved great success in video recognition, yet still struggles to recognize novel actions when faced with only a few examples. To tackle this challenge, few-shot action recognition methods have been proposed to transfer knowledge from a source dataset to a novel target dataset with only one or a few labeled videos. However, existing methods mainly focus on modeling the temporal relations between the query and support videos while ignoring the spatial relations. In this paper, we find that the spatial misalignment between objects also occurs in videos, notably more common than the temporal inconsistency. We are thus motivated to investigate the importance of spatial relations and propose a more accurate few-shot action recognition method that leverages both spatial and temporal information. Particularly, a novel Spatial Alignment Cross Transformer (SA-CT) which learns to re-adjust the spatial relations and incorporates the temporal information is contributed. Experiments reveal that, even without using any temporal information, the performance of SA-CT is comparable to temporal based methods on 3/4 benchmarks. To further incorporate the temporal information, we propose a simple yet effective Temporal Mixer module. The Temporal Mixer enhances the video representation and improves the performance of the full SA-CT model, achieving very competitive results. In this work, we also exploit large-scale pretrained models for few-shot action recognition, providing useful insights for this research direction.

Abstract: What will the future be? We wonder! In this survey, we explore the gap between current research in egocentric vision and the ever-anticipated future, where wearable computing, with outward facing cameras and digital overlays, is expected to be integrated in our every day lives. To understand this gap, the article starts by envisaging the future through character-based stories, showcasing through examples the limitations of current technology. We then provide a mapping between this future and previously defined research tasks. For each task, we survey its seminal works, current state-of-the-art methodologies and available datasets, then reflect on shortcomings that limit its applicability to future research. Note that this survey focuses on software models for egocentric vision, independent of any specific hardware. The paper concludes with recommendations for areas of immediate explorations so as to unlock our path to the future always-on, personalised and life-enhancing egocentric vision.

Abstract: Vision-Language Pretraining (VLP) has shown impressive results on diverse downstream tasks by offline training on large-scale datasets. Regarding the growing nature of real-world data, such an offline training paradigm on ever-expanding data is unsustainable, because models lack the continual learning ability to accumulate knowledge constantly. However, most continual learning studies are limited to uni-modal classification and existing multi-modal datasets cannot simulate continual non-stationary data stream scenarios. To support the study of Vision-Language Continual Pretraining (VLCP), we first contribute a comprehensive and unified benchmark dataset P9D which contains over one million product image-text pairs from 9 industries. The data from each industry as an independent task supports continual learning and conforms to the real-world long-tail nature to simulate pretraining on web data. We comprehensively study the characteristics and challenges of VLCP, and propose a new algorithm: Compatible momentum contrast with Topology Preservation, dubbed CTP. The compatible momentum model absorbs the knowledge of the current and previous-task models to flexibly update the modal feature. Moreover, Topology Preservation transfers the knowledge of embedding across tasks while preserving the flexibility of feature adjustment. The experimental results demonstrate our method not only achieves superior performance compared with other baselines but also does not bring an expensive training burden. Dataset and codes are available at https://github.com/KevinLight831/CTP.

Abstract: Cultural heritage applications and advanced machine learning models are creating a fruitful synergy to provide effective and accessible ways of interacting with artworks. Smart audio-guides, personalized art-related content and gamification approaches are just a few examples of how technology can be exploited to provide additional value to artists or exhibitions. Nonetheless, from a machine learning point of view, the amount of available artistic data is often not enough to train effective models. Off-the-shelf computer vision modules can still be exploited to some extent, yet a severe domain shift is present between art images and standard natural image datasets used to train such models. As a result, this can lead to degraded performance. This paper introduces a novel approach to address the challenges of limited annotated data and domain shifts in the cultural heritage domain. By leveraging generative vision-language models, we augment art datasets by generating diverse variations of artworks conditioned on their captions. This augmentation strategy enhances dataset diversity, bridging the gap between natural images and artworks, and improving the alignment of visual cues with knowledge from general-purpose datasets. The generated variations assist in training vision and language models with a deeper understanding of artistic characteristics and that are able to generate better captions with appropriate jargon.

Abstract: Accurate, robust, and real-time LiDAR-based odometry (LO) is imperative for many applications like robot navigation, globally consistent 3D scene map reconstruction, or safe motion-planning. Though LiDAR sensor is known for its precise range measurement, the non-uniform and uncertain point sampling density induce structural inconsistencies. Hence, existing supervised and unsupervised point set registration methods fail to establish one-to-one matching correspondences between LiDAR frames. We introduce a novel deep learning-based real-time (approx. 35-40ms per frame) LO method that jointly learns accurate frame-to-frame correspondences and model's predictive uncertainty (PU) as evidence to safe-guard LO predictions. In this work, we propose (i) partial optimal transportation of LiDAR feature descriptor for robust LO estimation, (ii) joint learning of predictive uncertainty while learning odometry over driving sequences, and (iii) demonstrate how PU can serve as evidence for necessary pose-graph optimization when LO network is either under or over confident. We evaluate our method on KITTI dataset and show competitive performance, even superior generalization ability over recent state-of-the-art approaches. Source codes are available.

Abstract: Sparse neural networks are a key factor in developing resource-efficient machine learning applications. We propose the novel and powerful sparse learning method Adaptive Regularized Training (ART) to compress dense into sparse networks. Instead of the commonly used binary mask during training to reduce the number of model weights, we inherently shrink weights close to zero in an iterative manner with increasing weight regularization. Our method compresses the pre-trained model knowledge into the weights of highest magnitude. Therefore, we introduce a novel regularization loss named HyperSparse that exploits the highest weights while conserving the ability of weight exploration. Extensive experiments on CIFAR and TinyImageNet show that our method leads to notable performance gains compared to other sparsification methods, especially in extremely high sparsity regimes up to 99.8 percent model sparsity. Additional investigations provide new insights into the patterns that are encoded in weights with high magnitudes.

Abstract: Continual shrinking of pattern dimensions in the semiconductor domain is making it increasingly difficult to inspect defects due to factors such as the presence of stochastic noise and the dynamic behavior of defect patterns and types. Conventional rule-based methods and non-parametric supervised machine learning algorithms like KNN mostly fail at the requirements of semiconductor defect inspection at these advanced nodes. Deep Learning (DL)-based methods have gained popularity in the semiconductor defect inspection domain because they have been proven robust towards these challenging scenarios. In this research work, we have presented an automated DL-based approach for efficient localization and classification of defects in SEM images. We have proposed SEMI-CenterNet (SEMI-CN), a customized CN architecture trained on SEM images of semiconductor wafer defects. The use of the proposed CN approach allows improved computational efficiency compared to previously studied DL models. SEMI-CN gets trained to output the center, class, size, and offset of a defect instance. This is different from the approach of most object detection models that use anchors for bounding box prediction. Previous methods predict redundant bounding boxes, most of which are discarded in postprocessing. CN mitigates this by only predicting boxes for likely defect center points. We train SEMI-CN on two datasets and benchmark two ResNet backbones for the framework. Initially, ResNet models pretrained on the COCO dataset undergo training using two datasets separately. Primarily, SEMI-CN shows significant improvement in inference time against previous research works. Finally, transfer learning (using weights of custom SEM dataset) is applied from ADI dataset to AEI dataset and vice-versa, which reduces the required training time for both backbones to reach the best mAP against conventional training method.

Abstract: Due to the flexible representation of arbitrary-shaped scene text and simple pipeline, bottom-up segmentation-based methods begin to be mainstream in real-time scene text detection. Despite great progress, these methods show deficiencies in robustness and still suffer from false positives and instance adhesion. Different from existing methods which integrate multiple-granularity features or multiple outputs, we resort to the perspective of representation learning in which auxiliary tasks are utilized to enable the encoder to jointly learn robust features with the main task of per-pixel classification during optimization. For semantic representation learning, we propose global-dense semantic contrast (GDSC), in which a vector is extracted for global semantic representation, then used to perform element-wise contrast with the dense grid features. To learn instance-aware representation, we propose to combine top-down modeling (TDM) with the bottom-up framework to provide implicit instance-level clues for the encoder. With the proposed GDSC and TDM, the encoder network learns stronger representation without introducing any parameters and computations during inference. Equipped with a very light decoder, the detector can achieve more robust real-time scene text detection. Experimental results on four public datasets show that the proposed method can outperform or be comparable to the state-of-the-art on both accuracy and speed. Specifically, the proposed method achieves 87.2% F-measure with 48.2 FPS on Total-Text and 89.6% F-measure with 36.9 FPS on MSRA-TD500 on a single GeForce RTX 2080 Ti GPU.

Abstract: Multiple object tracking (MOT) has been successfully investigated in computer vision. However, MOT for the videos captured by unmanned aerial vehicles (UAV) is still challenging due to small object size, blurred object appearance, and very large and/or irregular motion in both ground objects and UAV platforms. In this paper, we propose FOLT to mitigate these problems and reach fast and accurate MOT in UAV view. Aiming at speed-accuracy trade-off, FOLT adopts a modern detector and light-weight optical flow extractor to extract object detection features and motion features at a minimum cost. Given the extracted flow, the flow-guided feature augmentation is designed to augment the object detection feature based on its optical flow, which improves the detection of small objects. Then the flow-guided motion prediction is also proposed to predict the object's position in the next frame, which improves the tracking performance of objects with very large displacements between adjacent frames. Finally, the tracker matches the detected objects and predicted objects using a spatially matching scheme to generate tracks for every object. Experiments on Visdrone and UAVDT datasets show that our proposed model can successfully track small objects with large and irregular motion and outperform existing state-of-the-art methods in UAV-MOT tasks.

Abstract: Performance estimation under covariate shift is a crucial component of safe AI model deployment, especially for sensitive use-cases. Recently, several solutions were proposed to tackle this problem, most leveraging model predictions or softmax confidence to derive accuracy estimates. However, under dataset shifts, confidence scores may become ill-calibrated if samples are too far from the training distribution. In this work, we show that taking into account distances of test samples to their expected training distribution can significantly improve performance estimation under covariate shift. Precisely, we introduce a "distance-check" to flag samples that lie too far from the expected distribution, to avoid relying on their untrustworthy model outputs in the accuracy estimation step. We demonstrate the effectiveness of this method on 13 image classification tasks, across a wide-range of natural and synthetic distribution shifts and hundreds of models, with a median relative MAE improvement of 27% over the best baseline across all tasks, and SOTA performance on 10 out of 13 tasks. Our code is publicly available at https://github.com/melanibe/distance_matters_performance_estimation.

Abstract: Self-supervised monocular depth estimation methods typically rely on the reprojection error to capture geometric relationships between successive frames in static environments. However, this assumption does not hold in dynamic objects in scenarios, leading to errors during the view synthesis stage, such as feature mismatch and occlusion, which can significantly reduce the accuracy of the generated depth maps. To address this problem, we propose a novel dynamic cost volume that exploits residual optical flow to describe moving objects, improving incorrectly occluded regions in static cost volumes used in previous work. Nevertheless, the dynamic cost volume inevitably generates extra occlusions and noise, thus we alleviate this by designing a fusion module that makes static and dynamic cost volumes compensate for each other. In other words, occlusion from the static volume is refined by the dynamic volume, and incorrect information from the dynamic volume is eliminated by the static volume. Furthermore, we propose a pyramid distillation loss to reduce photometric error inaccuracy at low resolutions and an adaptive photometric error loss to alleviate the flow direction of the large gradient in the occlusion regions. We conducted extensive experiments on the KITTI and Cityscapes datasets, and the results demonstrate that our model outperforms previously published baselines for self-supervised monocular depth estimation.

Abstract: Blind face restoration aims at recovering high-quality face images from those with unknown degradations. Current algorithms mainly introduce priors to complement high-quality details and achieve impressive progress. However, most of these algorithms ignore abundant contextual information in the face and its interplay with the priors, leading to sub-optimal performance. Moreover, they pay less attention to the gap between the synthetic and real-world scenarios, limiting the robustness and generalization to real-world applications. In this work, we propose RestoreFormer++, which on the one hand introduces fully-spatial attention mechanisms to model the contextual information and the interplay with the priors, and on the other hand, explores an extending degrading model to help generate more realistic degraded face images to alleviate the synthetic-to-real-world gap. Compared with current algorithms, RestoreFormer++ has several crucial benefits. First, instead of using a multi-head self-attention mechanism like the traditional visual transformer, we introduce multi-head cross-attention over multi-scale features to fully explore spatial interactions between corrupted information and high-quality priors. In this way, it can facilitate RestoreFormer++ to restore face images with higher realness and fidelity. Second, in contrast to the recognition-oriented dictionary, we learn a reconstruction-oriented dictionary as priors, which contains more diverse high-quality facial details and better accords with the restoration target. Third, we introduce an extending degrading model that contains more realistic degraded scenarios for training data synthesizing, and thus helps to enhance the robustness and generalization of our RestoreFormer++ model. Extensive experiments show that RestoreFormer++ outperforms state-of-the-art algorithms on both synthetic and real-world datasets.

Abstract: In this paper, we draw inspiration from Alberto Elfes' pioneering work in 1989, where he introduced the concept of the occupancy grid as World Models for robots. We imbue the robot with a spatial-temporal world model, termed UniWorld, to perceive its surroundings and predict the future behavior of other participants. UniWorld involves initially predicting 4D geometric occupancy as the World Models for foundational stage and subsequently fine-tuning on downstream tasks. UniWorld can estimate missing information concerning the world state and predict plausible future states of the world. Besides, UniWorld's pre-training process is label-free, enabling the utilization of massive amounts of image-LiDAR pairs to build a Foundational Model.The proposed unified pre-training framework demonstrates promising results in key tasks such as motion prediction, multi-camera 3D object detection, and surrounding semantic scene completion. When compared to monocular pre-training methods on the nuScenes dataset, UniWorld shows a significant improvement of about 1.5% in IoU for motion prediction, 2.0% in mAP and 2.0% in NDS for multi-camera 3D object detection, as well as a 3% increase in mIoU for surrounding semantic scene completion. By adopting our unified pre-training method, a 25% reduction in 3D training annotation costs can be achieved, offering significant practical value for the implementation of real-world autonomous driving. Codes are publicly available at https://github.com/chaytonmin/UniWorld.

Abstract: In this paper, we present a new multi-branch neural network that simultaneously performs soft biometric (SB) prediction as an auxiliary modality and face recognition (FR) as the main task. Our proposed network named AAFace utilizes SB attributes to enhance the discriminative ability of FR representation. To achieve this goal, we propose an attribute-aware attentional integration (AAI) module to perform weighted integration of FR with SB feature maps. Our proposed AAI module is not only fully context-aware but also capable of learning complex relationships between input features by means of the sequential multi-scale channel and spatial sub-modules. Experimental results verify the superiority of our proposed network compared with the state-of-the-art (SoTA) SB prediction and FR methods.

Abstract: African penguins (Spheniscus demersus) are an endangered species. Little is known regarding their underwater hunting strategies and associated predation success rates, yet this is essential for guiding conservation. Modern bio-logging technology has the potential to provide valuable insights, but manually analysing large amounts of data from animal-borne video recorders (AVRs) is time-consuming. In this paper, we publish an animal-borne underwater video dataset of penguins and introduce a ready-to-deploy deep learning system capable of robustly detecting penguins ([email protected]%) and also instances of fish ([email protected]%). We note that the detectors benefit explicitly from air-bubble learning to improve accuracy. Extending this detector towards a dual-stream behaviour recognition network, we also provide the first results for identifying predation behaviour in penguin underwater videos. Whilst results are promising, further work is required for useful applicability of predation behaviour detection in field scenarios. In summary, we provide a highly reliable underwater penguin detector, a fish detector, and a valuable first attempt towards an automated visual detection of complex behaviours in a marine predator. We publish the networks, the DivingWithPenguins video dataset, annotations, splits, and weights for full reproducibility and immediate usability by practitioners.

Abstract: The works in literature classifying natural and computer generated images are mostly designed as binary tasks either considering natural images versus computer graphics images only or natural images versus GAN generated images only, but not natural images versus both classes of the generated images. Also, even though this forensic classification task of distinguishing natural and computer generated images gets the support of the new convolutional neural networks and transformer based architectures that can give remarkable classification accuracies, they are seen to fail over the images that have undergone some post-processing operations usually performed to deceive the forensic algorithms, such as JPEG compression, gaussian noise, etc. This work proposes a robust approach towards distinguishing natural and computer generated images including both, computer graphics and GAN generated images using a fusion of two vision transformers where each of the transformer networks operates in different color spaces, one in RGB and the other in YCbCr color space. The proposed approach achieves high performance gain when compared to a set of baselines, and also achieves higher robustness and generalizability than the baselines. The features of the proposed model when visualized are seen to obtain higher separability for the classes than the input image features and the baseline features. This work also studies the attention map visualizations of the networks of the fused model and observes that the proposed methodology can capture more image information relevant to the forensic task of classifying natural and generated images.

Abstract: Eye-tracking plays a crucial role in the development of virtual reality devices, neuroscience research, and psychology. Despite its significance in numerous applications, achieving an accurate, robust, and fast eye-tracking solution remains a considerable challenge for current state-of-the-art methods. While existing reflection-based techniques (e.g., "glint tracking") are considered the most accurate, their performance is limited by their reliance on sparse 3D surface data acquired solely from the cornea surface. In this paper, we rethink the way how specular reflections can be used for eye tracking: We propose a novel method for accurate and fast evaluation of the gaze direction that exploits teachings from single-shot phase-measuring-deflectometry (PMD). In contrast to state-of-the-art reflection-based methods, our method acquires dense 3D surface information of both cornea and sclera within only one single camera frame (single-shot). Improvements in acquired reflection surface points("glints") of factors $>3300 \times$ are easily achievable. We show the feasibility of our approach with experimentally evaluated gaze errors of only $\leq 0.25^\circ$ demonstrating a significant improvement over the current state-of-the-art.

Abstract: The synthesis of human motion has traditionally been addressed through task-dependent models that focus on specific challenges, such as predicting future motions or filling in intermediate poses conditioned on known key-poses. In this paper, we present a novel task-independent model called UNIMASK-M, which can effectively address these challenges using a unified architecture. Our model obtains comparable or better performance than the state-of-the-art in each field. Inspired by Vision Transformers (ViTs), our UNIMASK-M model decomposes a human pose into body parts to leverage the spatio-temporal relationships existing in human motion. Moreover, we reformulate various pose-conditioned motion synthesis tasks as a reconstruction problem with different masking patterns given as input. By explicitly informing our model about the masked joints, our UNIMASK-M becomes more robust to occlusions. Experimental results show that our model successfully forecasts human motion on the Human3.6M dataset. Moreover, it achieves state-of-the-art results in motion inbetweening on the LaFAN1 dataset, particularly in long transition periods. More information can be found on the project website https://sites.google.com/view/estevevallsmascaro/publications/unimask-m.

Abstract: In this paper, we study the problem of end-to-end multi-person pose estimation. State-of-the-art solutions adopt the DETR-like framework, and mainly develop the complex decoder, e.g., regarding pose estimation as keypoint box detection and combining with human detection in ED-Pose, hierarchically predicting with pose decoder and joint (keypoint) decoder in PETR. We present a simple yet effective transformer approach, named Group Pose. We simply regard $K$-keypoint pose estimation as predicting a set of $N\times K$ keypoint positions, each from a keypoint query, as well as representing each pose with an instance query for scoring $N$ pose predictions. Motivated by the intuition that the interaction, among across-instance queries of different types, is not directly helpful, we make a simple modification to decoder self-attention. We replace single self-attention over all the $N\times(K+1)$ queries with two subsequent group self-attentions: (i) $N$ within-instance self-attention, with each over $K$ keypoint queries and one instance query, and (ii) $(K+1)$ same-type across-instance self-attention, each over $N$ queries of the same type. The resulting decoder removes the interaction among across-instance type-different queries, easing the optimization and thus improving the performance. Experimental results on MS COCO and CrowdPose show that our approach without human box supervision is superior to previous methods with complex decoders, and even is slightly better than ED-Pose that uses human box supervision. $\href{https://github.com/Michel-liu/GroupPose-Paddle}{\rm Paddle}$ and $\href{https://github.com/Michel-liu/GroupPose}{\rm PyTorch}$ code are available.

Abstract: Restoring facial details from low-quality (LQ) images has remained a challenging problem due to its ill-posedness induced by various degradations in the wild. The existing codebook prior mitigates the ill-posedness by leveraging an autoencoder and learned codebook of high-quality (HQ) features, achieving remarkable quality. However, existing approaches in this paradigm frequently depend on a single encoder pre-trained on HQ data for restoring HQ images, disregarding the domain gap between LQ and HQ images. As a result, the encoding of LQ inputs may be insufficient, resulting in suboptimal performance. To tackle this problem, we propose a novel dual-branch framework named DAEFR. Our method introduces an auxiliary LQ branch that extracts crucial information from the LQ inputs. Additionally, we incorporate association training to promote effective synergy between the two branches, enhancing code prediction and output quality. We evaluate the effectiveness of DAEFR on both synthetic and real-world datasets, demonstrating its superior performance in restoring facial details.

Abstract: With a strong understanding of the target domain from natural language, we produce promising results in translating across large domain gaps and bringing skeletons back to life. In this work, we use text-guided latent diffusion models for zero-shot image-to-image translation (I2I) across large domain gaps (longI2I), where large amounts of new visual features and new geometry need to be generated to enter the target domain. Being able to perform translations across large domain gaps has a wide variety of real-world applications in criminology, astrology, environmental conservation, and paleontology. In this work, we introduce a new task Skull2Animal for translating between skulls and living animals. On this task, we find that unguided Generative Adversarial Networks (GANs) are not capable of translating across large domain gaps. Instead of these traditional I2I methods, we explore the use of guided diffusion and image editing models and provide a new benchmark model, Revive-2I, capable of performing zero-shot I2I via text-prompting latent diffusion models. We find that guidance is necessary for longI2I because, to bridge the large domain gap, prior knowledge about the target domain is needed. In addition, we find that prompting provides the best and most scalable information about the target domain as classifier-guided diffusion models require retraining for specific use cases and lack stronger constraints on the target domain because of the wide variety of images they are trained on.