Computer Vision and Pattern Recognition (cs.CV)

Abstract: In recent years, sketch-based 3D shape retrieval has attracted growing attention. While many previous studies have focused on cross-modal matching between hand-drawn sketches and 3D shapes, the critical issue of how to handle low-quality and noisy samples in sketch data has been largely neglected. This paper presents an uncertainty-aware cross-modal transfer network (UACTN) that addresses this issue. UACTN decouples the representation learning of sketches and 3D shapes into two separate tasks: classification-based sketch uncertainty learning and 3D shape feature transfer. We first introduce an end-to-end classification-based approach that simultaneously learns sketch features and uncertainty, allowing uncertainty to prevent overfitting noisy sketches by assigning different levels of importance to clean and noisy sketches. Then, 3D shape features are mapped into the pre-learned sketch embedding space for feature alignment. Extensive experiments and ablation studies on two benchmarks demonstrate the superiority of our proposed method compared to state-of-the-art methods.

Abstract: Human-object interaction (HOI) detection is an important part of understanding human activities and visual scenes. The long-tailed distribution of labeled instances is a primary challenge in HOI detection, promoting research in few-shot and zero-shot learning. Inspired by the combinatorial nature of HOI triplets, some existing approaches adopt the idea of compositional learning, in which object and action features are learned individually and re-composed as new training samples. However, these methods follow the CNN-based two-stage paradigm with limited feature extraction ability, and often rely on auxiliary information for better performance. Without introducing any additional information, we creatively propose a transformer-based framework for compositional HOI learning. Human-object pair representations and interaction representations are re-composed across different HOI instances, which involves richer contextual information and promotes the generalization of knowledge. Experiments show our simple but effective method achieves state-of-the-art performance, especially on rare HOI classes.

Abstract: Detecting dental diseases through panoramic X-rays images is a standard procedure for dentists. Normally, a dentist need to identify diseases and find the infected teeth. While numerous machine learning models adopting this two-step procedure have been developed, there has not been an end-to-end model that can identify teeth and their associated diseases at the same time. To fill the gap, we develop YOLOrtho, a unified framework for teeth enumeration and dental disease detection. We develop our model on Dentex Challenge 2023 data, which consists of three distinct types of annotated data. The first part is labeled with quadrant, and the second part is labeled with quadrant and enumeration and the third part is labeled with quadrant, enumeration and disease. To further improve detection, we make use of Tufts Dental public dataset. To fully utilize the data and learn both teeth detection and disease identification simultaneously, we formulate diseases as attributes attached to their corresponding teeth. Due to the nature of position relation in teeth enumeration, We replace convolution layer with CoordConv in our model to provide more position information for the model. We also adjust the model architecture and insert one more upsampling layer in FPN in favor of large object detection. Finally, we propose a post-process strategy for teeth layout that corrects teeth enumeration based on linear sum assignment. Results from experiments show that our model exceeds large Diffusion-based model.

Abstract: Most NeRF-based models are designed for learning the entire scene, and complex scenes can lead to longer learning times and poorer rendering effects. This paper utilizes scene semantic priors to make improvements in fast training, allowing the network to focus on the specific targets and not be affected by complex backgrounds. The training speed can be increased by 7.78 times with better rendering effect, and small to medium sized targets can be rendered faster. In addition, this improvement applies to all NeRF-based models. Considering the inherent multi-view consistency and smoothness of NeRF, this paper also studies weak supervision by sparsely sampling negative ray samples. With this method, training can be further accelerated and rendering quality can be maintained. Finally, this paper extends pixel semantic and color rendering formulas and proposes a new scene editing technique that can achieve unique displays of the specific semantic targets or masking them in rendering. To address the problem of unsupervised regions incorrect inferences in the scene, we also designed a self-supervised loop that combines morphological operations and clustering.

Abstract: This paper proposes a novel and physically interpretable method for face editing based on arbitrary text prompts. Different from previous GAN-inversion-based face editing methods that manipulate the latent space of GANs, or diffusion-based methods that model image manipulation as a reverse diffusion process, we regard the face editing process as imposing vector flow fields on face images, representing the offset of spatial coordinates and color for each image pixel. Under the above-proposed paradigm, we represent the vector flow field in two ways: 1) explicitly represent the flow vectors with rasterized tensors, and 2) implicitly parameterize the flow vectors as continuous, smooth, and resolution-agnostic neural fields, by leveraging the recent advances of implicit neural representations. The flow vectors are iteratively optimized under the guidance of the pre-trained Contrastive Language-Image Pretraining~(CLIP) model by maximizing the correlation between the edited image and the text prompt. We also propose a learning-based one-shot face editing framework, which is fast and adaptable to any text prompt input. Our method can also be flexibly extended to real-time video face editing. Compared with state-of-the-art text-driven face editing methods, our method can generate physically interpretable face editing results with high identity consistency and image quality. Our code will be made publicly available.

Abstract: Billions of people are sharing their daily live images on social media everyday. However, malicious collectors use deep face recognition systems to easily steal their biometric information (e.g., faces) from these images. Some studies are being conducted to generate encrypted face photos using adversarial attacks by introducing imperceptible perturbations to reduce face information leakage. However, existing studies need stronger black-box scenario feasibility and more natural visual appearances, which challenge the feasibility of privacy protection. To address these problems, we propose a frequency-restricted identity-agnostic (FRIA) framework to encrypt face images from unauthorized face recognition without access to personal information. As for the weak black-box scenario feasibility, we obverse that representations of the average feature in multiple face recognition models are similar, thus we propose to utilize the average feature via the crawled dataset from the Internet as the target to guide the generation, which is also agnostic to identities of unknown face recognition systems; in nature, the low-frequency perturbations are more visually perceptible by the human vision system. Inspired by this, we restrict the perturbation in the low-frequency facial regions by discrete cosine transform to achieve the visual naturalness guarantee. Extensive experiments on several face recognition models demonstrate that our FRIA outperforms other state-of-the-art methods in generating more natural encrypted faces while attaining high black-box attack success rates of 96%. In addition, we validate the efficacy of FRIA using real-world black-box commercial API, which reveals the potential of FRIA in practice. Our codes can be found in https://github.com/XinDong10/FRIA.

Abstract: Deploying 3D detectors in unfamiliar domains has been demonstrated to result in a drastic drop of up to 70-90% in detection rate due to variations in lidar, geographical region, or weather conditions from their original training dataset. This domain gap leads to missing detections for densely observed objects, misaligned confidence scores, and increased high-confidence false positives, rendering the detector highly unreliable. To address this, we introduce MS3D++, a self-training framework for multi-source unsupervised domain adaptation in 3D object detection. MS3D++ provides a straightforward approach to domain adaptation by generating high-quality pseudo-labels, enabling the adaptation of 3D detectors to a diverse range of lidar types, regardless of their density. Our approach effectively fuses predictions of an ensemble of multi-frame pre-trained detectors from different source domains to improve domain generalization. We subsequently refine the predictions temporally to ensure temporal consistency in box localization and object classification. Furthermore, we present an in-depth study into the performance and idiosyncrasies of various 3D detector components in a cross-domain context, providing valuable insights for improved cross-domain detector ensembling. Experimental results on Waymo, nuScenes and Lyft demonstrate that detectors trained with MS3D++ pseudo-labels achieve state-of-the-art performance, comparable to training with human-annotated labels in Bird's Eye View (BEV) evaluation for both low and high density lidar.

Abstract: Recent progress in weakly supervised object detection is featured by a combination of multiple instance detection networks (MIDN) and ordinal online refinement. However, with only image-level annotation, MIDN inevitably assigns high scores to some unexpected region proposals when generating pseudo labels. These inaccurate high-scoring region proposals will mislead the training of subsequent refinement modules and thus hamper the detection performance. In this work, we explore how to ameliorate the quality of pseudo-labeling in MIDN. Formally, we devise Cyclic-Bootstrap Labeling (CBL), a novel weakly supervised object detection pipeline, which optimizes MIDN with rank information from a reliable teacher network. Specifically, we obtain this teacher network by introducing a weighted exponential moving average strategy to take advantage of various refinement modules. A novel class-specific ranking distillation algorithm is proposed to leverage the output of weighted ensembled teacher network for distilling MIDN with rank information. As a result, MIDN is guided to assign higher scores to accurate proposals among their neighboring ones, thus benefiting the subsequent pseudo labeling. Extensive experiments on the prevalent PASCAL VOC 2007 \& 2012 and COCO datasets demonstrate the superior performance of our CBL framework. Code will be available at https://github.com/Yinyf0804/WSOD-CBL/.

Abstract: We study the image-based geolocalization problem that aims to locate ground-view query images on cartographic maps. Previous methods often utilize cross-view localization techniques to match ground-view query images with 2D maps. However, the performance of these methods is frequently unsatisfactory due to the significant cross-view appearance differences. In this paper, we extend cross-view matching to 2.5D spaces, where the heights of the structures - such as trees, buildings, and other objects - can provide additional information to guide the cross-view matching. We present a new approach to learning representative embeddings from multi-model data. Specifically, we first align 2D maps to ground-view panoramic images with polar transform to reduce the gap between panoramic images and maps. Then we leverage global fusion to fuse the multi-modal features from 2D and 2.5D maps to increase the distinctiveness of location embeddings. We construct the first large-scale ground-to-2.5D map geolocalization dataset to validate our method and facilitate the research. We test our learned embeddings on two popular localization approaches, i.e., single-image based localization, and route based localization. Extensive experiments demonstrate that our proposed method achieves significantly higher localization accuracy and faster convergence than previous 2D map-based approaches.

Abstract: The video grounding (VG) task aims to locate the queried action or event in an untrimmed video based on rich linguistic descriptions. Existing proposal-free methods are trapped in complex interaction between video and query, overemphasizing cross-modal feature fusion and feature correlation for VG. In this paper, we propose a novel boundary regression paradigm that performs regression token learning in a transformer. Particularly, we present a simple but effective proposal-free framework, namely Video Grounding Transformer (ViGT), which predicts the temporal boundary using a learnable regression token rather than multi-modal or cross-modal features. In ViGT, the benefits of a learnable token are manifested as follows. (1) The token is unrelated to the video or the query and avoids data bias toward the original video and query. (2) The token simultaneously performs global context aggregation from video and query features. First, we employed a sharing feature encoder to project both video and query into a joint feature space before performing cross-modal co-attention (i.e., video-to-query attention and query-to-video attention) to highlight discriminative features in each modality. Furthermore, we concatenated a learnable regression token [REG] with the video and query features as the input of a vision-language transformer. Finally, we utilized the token [REG] to predict the target moment and visual features to constrain the foreground and background probabilities at each timestamp. The proposed ViGT performed well on three public datasets: ANet Captions, TACoS and YouCookII. Extensive ablation studies and qualitative analysis further validated the interpretability of ViGT.

Abstract: Deep neural networks are vulnerable to universal adversarial perturbation (UAP), an instance-agnostic perturbation capable of fooling the target model for most samples. Compared to instance-specific adversarial examples, UAP is more challenging as it needs to generalize across various samples and models. In this paper, we examine the serious dilemma of UAP generation methods from a generalization perspective -- the gradient vanishing problem using small-batch stochastic gradient optimization and the local optima problem using large-batch optimization. To address these problems, we propose a simple and effective method called Stochastic Gradient Aggregation (SGA), which alleviates the gradient vanishing and escapes from poor local optima at the same time. Specifically, SGA employs the small-batch training to perform multiple iterations of inner pre-search. Then, all the inner gradients are aggregated as a one-step gradient estimation to enhance the gradient stability and reduce quantization errors. Extensive experiments on the standard ImageNet dataset demonstrate that our method significantly enhances the generalization ability of UAP and outperforms other state-of-the-art methods. The code is available at https://github.com/liuxuannan/Stochastic-Gradient-Aggregation.

Abstract: Convolutional Neural Networks (CNN) have become a common choice for industrial quality control, as well as other critical applications in the Industry 4.0. When these CNNs behave in ways unexpected to human users or developers, severe consequences can arise, such as economic losses or an increased risk to human life. Concept extraction techniques can be applied to increase the reliability and transparency of CNNs through generating global explanations for trained neural network models. The decisive features of image datasets in quality control often depend on the feature's scale; for example, the size of a hole or an edge. However, existing concept extraction methods do not correctly represent scale, which leads to problems interpreting these models as we show herein. To address this issue, we introduce the Scale-Preserving Automatic Concept Extraction (SPACE) algorithm, as a state-of-the-art alternative concept extraction technique for CNNs, focused on industrial applications. SPACE is specifically designed to overcome the aforementioned problems by avoiding scale changes throughout the concept extraction process. SPACE proposes an approach based on square slices of input images, which are selected and then tiled before being clustered into concepts. Our method provides explanations of the models' decision-making process in the form of human-understandable concepts. We evaluate SPACE on three image classification datasets in the context of industrial quality control. Through experimental results, we illustrate how SPACE outperforms other methods and provides actionable insights on the decision mechanisms of CNNs. Finally, code for the implementation of SPACE is provided.

Abstract: Efficient RGB-D semantic segmentation has received considerable attention in mobile robots, which plays a vital role in analyzing and recognizing environmental information. According to previous studies, depth information can provide corresponding geometric relationships for objects and scenes, but actual depth data usually exist as noise. To avoid unfavorable effects on segmentation accuracy and computation, it is necessary to design an efficient framework to leverage cross-modal correlations and complementary cues. In this paper, we propose an efficient lightweight encoder-decoder network that reduces the computational parameters and guarantees the robustness of the algorithm. Working with channel and spatial fusion attention modules, our network effectively captures multi-level RGB-D features. A globally guided local affinity context module is proposed to obtain sufficient high-level context information. The decoder utilizes a lightweight residual unit that combines short- and long-distance information with a few redundant computations. Experimental results on NYUv2, SUN RGB-D, and Cityscapes datasets show that our method achieves a better trade-off among segmentation accuracy, inference time, and parameters than the state-of-the-art methods. The source code will be at https://github.com/MVME-HBUT/SGACNet

Abstract: Text-to-image synthesis has achieved high-quality results with recent advances in diffusion models. However, text input alone has high spatial ambiguity and limited user controllability. Most existing methods allow spatial control through additional visual guidance (e.g, sketches and semantic masks) but require additional training with annotated images. In this paper, we propose a method for spatially controlling text-to-image generation without further training of diffusion models. Our method is based on the insight that the cross-attention maps reflect the positional relationship between words and pixels. Our aim is to control the attention maps according to given semantic masks and text prompts. To this end, we first explore a simple approach of directly swapping the cross-attention maps with constant maps computed from the semantic regions. Moreover, we propose masked-attention guidance, which can generate images more faithful to semantic masks than the first approach. Masked-attention guidance indirectly controls attention to each word and pixel according to the semantic regions by manipulating noise images fed to diffusion models. Experiments show that our method enables more accurate spatial control than baselines qualitatively and quantitatively.

Abstract: Benefiting from prompt tuning, recent years have witnessed the promising performance of pre-trained vision-language models, e.g., CLIP, on versatile downstream tasks. In this paper, we focus on a particular setting of learning adaptive prompts on the fly for each test sample from an unseen new domain, which is known as test-time prompt tuning (TPT). Existing TPT methods typically rely on data augmentation and confidence selection. However, conventional data augmentation techniques, e.g., random resized crops, suffers from the lack of data diversity, while entropy-based confidence selection alone is not sufficient to guarantee prediction fidelity. To address these issues, we propose a novel TPT method, named DiffTPT, which leverages pre-trained diffusion models to generate diverse and informative new data. Specifically, we incorporate augmented data by both conventional method and pre-trained stable diffusion to exploit their respective merits, improving the models ability to adapt to unknown new test data. Moreover, to ensure the prediction fidelity of generated data, we introduce a cosine similarity-based filtration technique to select the generated data with higher similarity to the single test sample. Our experiments on test datasets with distribution shifts and unseen categories demonstrate that DiffTPT improves the zero-shot accuracy by an average of 5.13\% compared to the state-of-the-art TPT method. Our code and models will be publicly released.

Abstract: Computing platforms in autonomous vehicles record large amounts of data from many sensors, process the data through machine learning models, and make decisions to ensure the vehicle's safe operation. Fast, accurate, and reliable decision-making is critical. Traditional computer processors lack the power and flexibility needed for the perception and machine vision demands of advanced autonomous driving tasks. Hardware accelerators are special-purpose coprocessors that help autonomous vehicles meet performance requirements for higher levels of autonomy. This paper provides an overview of ML accelerators with examples of their use for machine vision in autonomous vehicles. We offer recommendations for researchers and practitioners and highlight a trajectory for ongoing and future research in this emerging field.

Abstract: This paper presents a computer-aided cytology diagnosis system designed for animals, focusing on image quality assessment (IQA) using Convolutional Neural Networks (CNNs). The system's building blocks are tailored to seamlessly integrate IQA, ensuring reliable performance in disease classification. We extensively investigate the CNN's ability to handle various image variations and scenarios, analyzing the impact on detecting low-quality input data. Additionally, the network's capacity to differentiate valid cellular samples from those with artifacts is evaluated. Our study employs a ResNet18 network architecture and explores the effects of input sizes and cropping strategies on model performance. The research sheds light on the significance of CNN-based IQA in computer-aided cytology diagnosis for animals, enhancing the accuracy of disease classification.

Abstract: Denoising Diffusion Models (DDM) are emerging as the cutting-edge technology in the realm of deep generative modeling, challenging the dominance of Generative Adversarial Networks. However, effectively exploring the latent space's semantics and identifying compelling trajectories for manipulating and editing important attributes of the generated samples remains challenging, primarily due to the high-dimensional nature of the latent space. In this study, we specifically concentrate on face rotation, which is known to be one of the most intricate editing operations. By leveraging a recent embedding technique for Denoising Diffusion Implicit Models (DDIM), we achieve, in many cases, noteworthy manipulations encompassing a wide rotation angle of $\pm 30^o$, preserving the distinct characteristics of the individual. Our methodology exploits the computation of trajectories approximating clouds of latent representations of dataset samples with different yaw rotations through linear regression. Specific trajectories are obtained by restricting the analysis to subsets of data sharing significant attributes with the source image. One of these attributes is the light provenance: a byproduct of our research is a labeling of CelebA, categorizing images into three major groups based on the illumination direction: left, center, and right.

Abstract: In monocular depth estimation, uncertainty estimation approaches mainly target the data uncertainty introduced by image noise. In contrast to prior work, we address the uncertainty due to lack of knowledge, which is relevant for the detection of data not represented by the training distribution, the so-called out-of-distribution (OOD) data. Motivated by anomaly detection, we propose to detect OOD images from an encoder-decoder depth estimation model based on the reconstruction error. Given the features extracted with the fixed depth encoder, we train an image decoder for image reconstruction using only in-distribution data. Consequently, OOD images result in a high reconstruction error, which we use to distinguish between in- and out-of-distribution samples. We built our experiments on the standard NYU Depth V2 and KITTI benchmarks as in-distribution data. Our post hoc method performs astonishingly well on different models and outperforms existing uncertainty estimation approaches without modifying the trained encoder-decoder depth estimation model.

Abstract: Creating realistic 3D facial animation is crucial for various applications in the movie production and gaming industry, especially with the burgeoning demand in the metaverse. However, prevalent methods such as blendshape-based approaches and facial rigging techniques are time-consuming, labor-intensive, and lack standardized configurations, making facial animation production challenging and costly. In this paper, we propose a novel self-supervised framework, Versatile Face Animator, which combines facial motion capture with motion retargeting in an end-to-end manner, eliminating the need for blendshapes or rigs. Our method has the following two main characteristics: 1) we propose an RGBD animation module to learn facial motion from raw RGBD videos by hierarchical motion dictionaries and animate RGBD images rendered from 3D facial mesh coarse-to-fine, enabling facial animation on arbitrary 3D characters regardless of their topology, textures, blendshapes, and rigs; and 2) we introduce a mesh retarget module to utilize RGBD animation to create 3D facial animation by manipulating facial mesh with controller transformations, which are estimated from dense optical flow fields and blended together with geodesic-distance-based weights. Comprehensive experiments demonstrate the effectiveness of our proposed framework in generating impressive 3D facial animation results, highlighting its potential as a promising solution for the cost-effective and efficient production of facial animation in the metaverse.

Abstract: Structural re-parameterization is a general training scheme for Convolutional Neural Networks (CNNs), which achieves performance improvement without increasing inference cost. As Vision Transformers (ViTs) are gradually surpassing CNNs in various visual tasks, one may question: if a training scheme specifically for ViTs exists that can also achieve performance improvement without increasing inference cost? Recently, Mixture-of-Experts (MoE) has attracted increasing attention, as it can efficiently scale up the capacity of Transformers at a fixed cost through sparsely activated experts. Considering that MoE can also be viewed as a multi-branch structure, can we utilize MoE to implement a ViT training scheme similar to structural re-parameterization? In this paper, we affirmatively answer these questions, with a new general training strategy for ViTs. Specifically, we decouple the training and inference phases of ViTs. During training, we replace some Feed-Forward Networks (FFNs) of the ViT with specially designed, more efficient MoEs that assign tokens to experts by random uniform partition, and perform Experts Weights Averaging (EWA) on these MoEs at the end of each iteration. After training, we convert each MoE into an FFN by averaging the experts, transforming the model back into original ViT for inference. We further provide a theoretical analysis to show why and how it works. Comprehensive experiments across various 2D and 3D visual tasks, ViT architectures, and datasets validate the effectiveness and generalizability of the proposed training scheme. Besides, our training scheme can also be applied to improve performance when fine-tuning ViTs. Lastly, but equally important, the proposed EWA technique can significantly improve the effectiveness of naive MoE in various 2D visual small datasets and 3D visual tasks.

Abstract: GAN inversion is indispensable for applying the powerful editability of GAN to real images. However, existing methods invert video frames individually often leading to undesired inconsistent results over time. In this paper, we propose a unified recurrent framework, named \textbf{R}ecurrent v\textbf{I}deo \textbf{G}AN \textbf{I}nversion and e\textbf{D}iting (RIGID), to explicitly and simultaneously enforce temporally coherent GAN inversion and facial editing of real videos. Our approach models the temporal relations between current and previous frames from three aspects. To enable a faithful real video reconstruction, we first maximize the inversion fidelity and consistency by learning a temporal compensated latent code. Second, we observe incoherent noises lie in the high-frequency domain that can be disentangled from the latent space. Third, to remove the inconsistency after attribute manipulation, we propose an \textit{in-between frame composition constraint} such that the arbitrary frame must be a direct composite of its neighboring frames. Our unified framework learns the inherent coherence between input frames in an end-to-end manner, and therefore it is agnostic to a specific attribute and can be applied to arbitrary editing of the same video without re-training. Extensive experiments demonstrate that RIGID outperforms state-of-the-art methods qualitatively and quantitatively in both inversion and editing tasks. The deliverables can be found in \url{https://cnnlstm.github.io/RIGID}

Abstract: Time-space diagrams are essential tools for analyzing traffic patterns and optimizing transportation infrastructure and traffic management strategies. Traditional data collection methods for these diagrams have limitations in terms of temporal and spatial coverage. Recent advancements in camera technology have overcome these limitations and provided extensive urban data. In this study, we propose an innovative approach to constructing time-space diagrams by utilizing street-view video sequences captured by cameras mounted on moving vehicles. Using the state-of-the-art YOLOv5, StrongSORT, and photogrammetry techniques for distance calculation, we can infer vehicle trajectories from the video data and generate time-space diagrams. To evaluate the effectiveness of our proposed method, we utilized datasets from the KITTI computer vision benchmark suite. The evaluation results demonstrate that our approach can generate trajectories from video data, although there are some errors that can be mitigated by improving the performance of the detector, tracker, and distance calculation components. In conclusion, the utilization of street-view video sequences captured by cameras mounted on moving vehicles, combined with state-of-the-art computer vision techniques, has immense potential for constructing comprehensive time-space diagrams. These diagrams offer valuable insights into traffic patterns and contribute to the design of transportation infrastructure and traffic management strategies.

Abstract: Latest methods for visual counterfactual explanations (VCE) harness the power of deep generative models to synthesize new examples of high-dimensional images of impressive quality. However, it is currently difficult to compare the performance of these VCE methods as the evaluation procedures largely vary and often boil down to visual inspection of individual examples and small scale user studies. In this work, we propose a framework for systematic, quantitative evaluation of the VCE methods and a minimal set of metrics to be used. We use this framework to explore the effects of certain crucial design choices in the latest diffusion-based generative models for VCEs of natural image classification (ImageNet). We conduct a battery of ablation-like experiments, generating thousands of VCEs for a suite of classifiers of various complexity, accuracy and robustness. Our findings suggest multiple directions for future advancements and improvements of VCE methods. By sharing our methodology and our approach to tackle the computational challenges of such a study on a limited hardware setup (including the complete code base), we offer a valuable guidance for researchers in the field fostering consistency and transparency in the assessment of counterfactual explanations.

Abstract: Virtual try-on is a critical image synthesis task that aims to transfer clothes from one image to another while preserving the details of both humans and clothes. While many existing methods rely on Generative Adversarial Networks (GANs) to achieve this, flaws can still occur, particularly at high resolutions. Recently, the diffusion model has emerged as a promising alternative for generating high-quality images in various applications. However, simply using clothes as a condition for guiding the diffusion model to inpaint is insufficient to maintain the details of the clothes. To overcome this challenge, we propose an exemplar-based inpainting approach that leverages a warping module to guide the diffusion model's generation effectively. The warping module performs initial processing on the clothes, which helps to preserve the local details of the clothes. We then combine the warped clothes with clothes-agnostic person image and add noise as the input of diffusion model. Additionally, the warped clothes is used as local conditions for each denoising process to ensure that the resulting output retains as much detail as possible. Our approach, namely Diffusion-based Conditional Inpainting for Virtual Try-ON (DCI-VTON), effectively utilizes the power of the diffusion model, and the incorporation of the warping module helps to produce high-quality and realistic virtual try-on results. Experimental results on VITON-HD demonstrate the effectiveness and superiority of our method.

Abstract: Despite the strong predictive performance of deep learning models for traffic prediction, their widespread deployment in real-world intelligent transportation systems has been restrained by a lack of interpretability. Uncertainty quantification (UQ) methods provide an approach to induce probabilistic reasoning, improve decision-making and enhance model deployment potential. To gain a comprehensive picture of the usefulness of existing UQ methods for traffic prediction and the relation between obtained uncertainties and city-wide traffic dynamics, we investigate their application to a large-scale image-based traffic dataset spanning multiple cities and time periods. We compare two epistemic and two aleatoric UQ methods on both temporal and spatio-temporal transfer tasks, and find that meaningful uncertainty estimates can be recovered. We further demonstrate how uncertainty estimates can be employed for unsupervised outlier detection on changes in city traffic dynamics. We find that our approach can capture both temporal and spatial effects on traffic behaviour in a representative case study for the city of Moscow. Our work presents a further step towards boosting uncertainty awareness in traffic prediction tasks, and aims to highlight the value contribution of UQ methods to a better understanding of city traffic dynamics.

Abstract: Automatic localization of text-lines in handwritten documents is still an open and challenging research problem. Various writing issues such as uneven spacing between the lines, oscillating and touching text, and the presence of skew become much more challenging when the case of complex handwritten document images are considered for segmentation directly in their respective compressed representation. This is because, the conventional way of processing compressed documents is through decompression, but here in this paper, we propose an idea that employs deep feature learning directly from the JPEG compressed coefficients without full decompression to accomplish text-line localization in the JPEG compressed domain. A modified U-Net architecture known as Compressed Text-Line Localization Network (CompTLL-UNet) is designed to accomplish it. The model is trained and tested with JPEG compressed version of benchmark datasets including ICDAR2017 (cBAD) and ICDAR2019 (cBAD), reporting the state-of-the-art performance with reduced storage and computational costs in the JPEG compressed domain.

Abstract: Driven by the increasing number of marine data science applications, there is a growing interest in surveying and exploring the vast, uncharted terrain of the deep sea with robotic platforms. Despite impressive results achieved by many on-land visual mapping algorithms in the past decades, transferring these methods from land to the deep sea remains a challenge due to harsh environmental conditions. Typically, deep-sea exploration involves the use of autonomous underwater vehicles (AUVs) equipped with high-resolution cameras and artificial illumination systems. However, images obtained in this manner often suffer from heterogeneous illumination and quality degradation due to attenuation and scattering, on top of refraction of light rays. All of this together often lets on-land SLAM approaches fail underwater or makes Structure-from-Motion approaches drift or omit difficult images, resulting in gaps, jumps or weakly registered areas. In this work, we present a system that incorporates recent developments in underwater imaging and visual mapping to facilitate automated robotic 3D reconstruction of hectares of seafloor. Our approach is efficient in that it detects and reconsiders difficult, weakly registered areas, to avoid omitting images and to make better use of limited dive time; on the other hand it is computationally efficient; leveraging a hybrid approach combining benefits from SLAM and Structure-from-Motion that runs much faster than incremental reconstructions while achieving at least on-par performance. The proposed system has been extensively tested and evaluated during several research cruises, demonstrating its robustness and practicality in real-world conditions.

Abstract: Current deep networks are very data-hungry and benefit from training on largescale datasets, which are often time-consuming to collect and annotate. By contrast, synthetic data can be generated infinitely using generative models such as DALL-E and diffusion models, with minimal effort and cost. In this paper, we present DatasetDM, a generic dataset generation model that can produce diverse synthetic images and the corresponding high-quality perception annotations (e.g., segmentation masks, and depth). Our method builds upon the pre-trained diffusion model and extends text-guided image synthesis to perception data generation. We show that the rich latent code of the diffusion model can be effectively decoded as accurate perception annotations using a decoder module. Training the decoder only needs less than 1% (around 100 images) manually labeled images, enabling the generation of an infinitely large annotated dataset. Then these synthetic data can be used for training various perception models for downstream tasks. To showcase the power of the proposed approach, we generate datasets with rich dense pixel-wise labels for a wide range of downstream tasks, including semantic segmentation, instance segmentation, and depth estimation. Notably, it achieves 1) state-of-the-art results on semantic segmentation and instance segmentation; 2) significantly more robust on domain generalization than using the real data alone; and state-of-the-art results in zero-shot segmentation setting; and 3) flexibility for efficient application and novel task composition (e.g., image editing). The project website and code can be found at https://weijiawu.github.io/DatasetDM_page/ and https://github.com/showlab/DatasetDM, respectively

Abstract: Weakly supervised object localization (WSOL) is one of the most popular and challenging tasks in computer vision. This task is to localize the objects in the images given only the image-level supervision. Recently, dividing WSOL into two parts (class-agnostic object localization and object classification) has become the state-of-the-art pipeline for this task. However, existing solutions under this pipeline usually suffer from the following drawbacks: 1) they are not flexible since they can only localize one object for each image due to the adopted single-class regression (SCR) for localization; 2) the generated pseudo bounding boxes may be noisy, but the negative impact of such noise is not well addressed. To remedy these drawbacks, we first propose to replace SCR with a binary-class detector (BCD) for localizing multiple objects, where the detector is trained by discriminating the foreground and background. Then we design a weighted entropy (WE) loss using the unlabeled data to reduce the negative impact of noisy bounding boxes. Extensive experiments on the popular CUB-200-2011 and ImageNet-1K datasets demonstrate the effectiveness of our method.

Abstract: Complex emotion recognition is a cognitive task that has so far eluded the same excellent performance of other tasks that are at or above the level of human cognition. Emotion recognition through facial expressions is particularly difficult due to the complexity of emotions expressed by the human face. For a machine to approach the same level of performance in this domain as a human, it may need to synthesise knowledge and understand new concepts in real-time as humans do. Humans are able to learn new concepts using only few examples, by distilling the important information from memories and discarding the rest. Similarly, continual learning methods learn new classes whilst retaining the knowledge of known classes, whilst few-shot learning methods are able to learn new classes using very few training examples. We propose a novel continual learning method inspired by human cognition and learning that can accurately recognise new compound expression classes using few training samples, by building on and retaining its knowledge of basic expression classes. Using GradCAM visualisations, we demonstrate the relationship between basic and compound facial expressions, which our method leverages through knowledge distillation and a novel Predictive Sorting Memory Replay. Our method achieves the current state-of-the-art in continual learning for complex facial expression recognition with 74.28% Overall Accuracy on new classes. We also demonstrate that using continual learning for complex facial expression recognition achieves far better performance than non-continual learning methods, improving on state-of-the-art non-continual learning methods by 13.95%. To the best of our knowledge, our work is also the first to apply few-shot learning to complex facial expression recognition, achieving the state-of-the-art with 100% accuracy using a single training sample for each expression class.

Abstract: The unprecedented photorealistic results achieved by recent text-to-image generative systems and their increasing use as plug-and-play content creation solutions make it crucial to understand their potential biases. In this work, we introduce three indicators to evaluate the realism, diversity and prompt-generation consistency of text-to-image generative systems when prompted to generate objects from across the world. Our indicators complement qualitative analysis of the broader impact of such systems by enabling automatic and efficient benchmarking of geographic disparities, an important step towards building responsible visual content creation systems. We use our proposed indicators to analyze potential geographic biases in state-of-the-art visual content creation systems and find that: (1) models have less realism and diversity of generations when prompting for Africa and West Asia than Europe, (2) prompting with geographic information comes at a cost to prompt-consistency and diversity of generated images, and (3) models exhibit more region-level disparities for some objects than others. Perhaps most interestingly, our indicators suggest that progress in image generation quality has come at the cost of real-world geographic representation. Our comprehensive evaluation constitutes a crucial step towards ensuring a positive experience of visual content creation for everyone.

Abstract: Recently, the DETR framework has emerged as the dominant approach for human--object interaction (HOI) research. In particular, two-stage transformer-based HOI detectors are amongst the most performant and training-efficient approaches. However, these often condition HOI classification on object features that lack fine-grained contextual information, eschewing pose and orientation information in favour of visual cues about object identity and box extremities. This naturally hinders the recognition of complex or ambiguous interactions. In this work, we study these issues through visualisations and carefully designed experiments. Accordingly, we investigate how best to re-introduce image features via cross-attention. With an improved query design, extensive exploration of keys and values, and box pair positional embeddings as spatial guidance, our model with enhanced predicate visual context (PViC) outperforms state-of-the-art methods on the HICO-DET and V-COCO benchmarks, while maintaining low training cost.

Abstract: Face forgery techniques have advanced rapidly and pose serious security threats. Existing face forgery detection methods try to learn generalizable features, but they still fall short of practical application. Additionally, finetuning these methods on historical training data is resource-intensive in terms of time and storage. In this paper, we focus on a novel and challenging problem: Continual Face Forgery Detection (CFFD), which aims to efficiently learn from new forgery attacks without forgetting previous ones. Specifically, we propose a Historical Distribution Preserving (HDP) framework that reserves and preserves the distributions of historical faces. To achieve this, we use universal adversarial perturbation (UAP) to simulate historical forgery distribution, and knowledge distillation to maintain the distribution variation of real faces across different models. We also construct a new benchmark for CFFD with three evaluation protocols. Our extensive experiments on the benchmarks show that our method outperforms the state-of-the-art competitors.

Abstract: The field of explainable artificial intelligence (XAI) aims to uncover the inner workings of complex deep neural models. While being crucial for safety-critical domains, XAI inherently lacks ground-truth explanations, making its automatic evaluation an unsolved problem. We address this challenge by proposing a novel synthetic vision dataset, named FunnyBirds, and accompanying automatic evaluation protocols. Our dataset allows performing semantically meaningful image interventions, e.g., removing individual object parts, which has three important implications. First, it enables analyzing explanations on a part level, which is closer to human comprehension than existing methods that evaluate on a pixel level. Second, by comparing the model output for inputs with removed parts, we can estimate ground-truth part importances that should be reflected in the explanations. Third, by mapping individual explanations into a common space of part importances, we can analyze a variety of different explanation types in a single common framework. Using our tools, we report results for 24 different combinations of neural models and XAI methods, demonstrating the strengths and weaknesses of the assessed methods in a fully automatic and systematic manner.