Computer Vision and Pattern Recognition (cs.CV)

Abstract: Visual question answering (VQA) is a Multidisciplinary research problem that pursued through practices of natural language processing and computer vision. Visual question answering automatically answers natural language questions according to the content of an image. Some testing questions require external knowledge to derive a solution. Such knowledge-based VQA uses various methods to retrieve features of image and text, and combine them to generate the answer. To generate knowledgebased answers either question dependent or image dependent knowledge retrieval methods are used. If knowledge about all the objects in the image is derived, then not all knowledge is relevant to the question. On other side only question related knowledge may lead to incorrect answers and over trained model that answers question that is irrelevant to image. Our proposed method takes image attributes and question features as input for knowledge derivation module and retrieves only question relevant knowledge about image objects which can provide accurate answers.

Abstract: In the domain of remote sensing image interpretation, road extraction from high-resolution aerial imagery has already been a hot research topic. Although deep CNNs have presented excellent results for semantic segmentation, the efficiency and capabilities of vision transformers are yet to be fully researched. As such, for accurate road extraction, a deep semantic segmentation neural network that utilizes the abilities of residual learning, HetConvs, UNet, and vision transformers, which is called \texttt{ResUNetFormer}, is proposed in this letter. The developed \texttt{ResUNetFormer} is evaluated on various cutting-edge deep learning-based road extraction techniques on the public Massachusetts road dataset. Statistical and visual results demonstrate the superiority of the \texttt{ResUNetFormer} over the state-of-the-art CNNs and vision transformers for segmentation. The code will be made available publicly at \url{https://github.com/aj1365/ResUNetFormer}.

Abstract: It is well-known that modern computer vision systems often exhibit behaviors misaligned with those of humans: from adversarial attacks to image corruptions, deep learning vision models suffer in a variety of settings that humans capably handle. In light of these phenomena, here we introduce another, orthogonal perspective studying the human-machine vision gap. We revisit the task of recovering images under degradation, first introduced over 30 years ago in the Recognition-by-Components theory of human vision. Specifically, we study the performance and behavior of neural networks on the seemingly simple task of classifying regular polygons at varying orders of degradation along their perimeters. To this end, we implement the Automated Shape Recoverability Test for rapidly generating large-scale datasets of perimeter-degraded regular polygons, modernizing the historically manual creation of image recoverability experiments. We then investigate the capacity of neural networks to recognize and recover such degraded shapes when initialized with different priors. Ultimately, we find that neural networks' behavior on this simple task conflicts with human behavior, raising a fundamental question of the robustness and learning capabilities of modern computer vision models.

Abstract: Reenacting facial images is an important task that can find numerous applications. We proposed IFaceUV, a fully differentiable pipeline that properly combines 2D and 3D information to conduct the facial reenactment task. The three-dimensional morphable face models (3DMMs) and corresponding UV maps are utilized to intuitively control facial motions and textures, respectively. Two-dimensional techniques based on 2D image warping is further required to compensate for missing components of the 3DMMs such as backgrounds, ear, hair and etc. In our pipeline, we first extract 3DMM parameters and corresponding UV maps from source and target images. Then, initial UV maps are refined by the UV map refinement network and it is rendered to the image with the motion manipulated 3DMM parameters. In parallel, we warp the source image according to the 2D flow field obtained from the 2D warping network. Rendered and warped images are combined in the final editing network to generate the final reenactment image. Additionally, we tested our model for the audio-driven facial reenactment task. Extensive qualitative and quantitative experiments illustrate the remarkable performance of our method compared to other state-of-the-art methods.

Abstract: We present a novel multi-view implicit surface reconstruction technique, termed StreetSurf, that is readily applicable to street view images in widely-used autonomous driving datasets, such as Waymo-perception sequences, without necessarily requiring LiDAR data. As neural rendering research expands rapidly, its integration into street views has started to draw interests. Existing approaches on street views either mainly focus on novel view synthesis with little exploration of the scene geometry, or rely heavily on dense LiDAR data when investigating reconstruction. Neither of them investigates multi-view implicit surface reconstruction, especially under settings without LiDAR data. Our method extends prior object-centric neural surface reconstruction techniques to address the unique challenges posed by the unbounded street views that are captured with non-object-centric, long and narrow camera trajectories. We delimit the unbounded space into three parts, close-range, distant-view and sky, with aligned cuboid boundaries, and adapt cuboid/hyper-cuboid hash-grids along with road-surface initialization scheme for finer and disentangled representation. To further address the geometric errors arising from textureless regions and insufficient viewing angles, we adopt geometric priors that are estimated using general purpose monocular models. Coupled with our implementation of efficient and fine-grained multi-stage ray marching strategy, we achieve state of the art reconstruction quality in both geometry and appearance within only one to two hours of training time with a single RTX3090 GPU for each street view sequence. Furthermore, we demonstrate that the reconstructed implicit surfaces have rich potential for various downstream tasks, including ray tracing and LiDAR simulation.

Abstract: Diffusion models have achieved impressive results in generating diverse and realistic data by employing multi-step denoising processes. However, the need for accommodating significant variations in input noise at each time-step has led to diffusion models requiring a large number of parameters for their denoisers. We have observed that diffusion models effectively act as filters for different frequency ranges at each time-step noise. While some previous works have introduced multi-expert strategies, assigning denoisers to different noise intervals, they overlook the importance of specialized operations for high and low frequencies. For instance, self-attention operations are effective at handling low-frequency components (low-pass filters), while convolutions excel at capturing high-frequency features (high-pass filters). In other words, existing diffusion models employ denoisers with the same architecture, without considering the optimal operations for each time-step noise. To address this limitation, we propose a novel approach called Multi-architecturE Multi-Expert (MEME), which consists of multiple experts with specialized architectures tailored to the operations required at each time-step interval. Through extensive experiments, we demonstrate that MEME outperforms large competitors in terms of both generation performance and computational efficiency.

Abstract: With the development of the multi-media internet, visual characteristics have become an important factor affecting user interests. Thus, incorporating visual features is a promising direction for further performance improvements in click-through rate (CTR) prediction. However, we found that simply injecting the image embeddings trained with established pre-training methods only has marginal improvements. We attribute the failure to two reasons: First, The pre-training methods are designed for well-defined computer vision tasks concentrating on semantic features, and they cannot learn personalized interest in recommendations. Secondly, pre-trained image embeddings only containing semantic information have little information gain, considering we already have semantic features such as categories and item titles as inputs in the CTR prediction task. We argue that a pre-training method tailored for recommendation is necessary for further improvements. To this end, we propose a recommendation-aware image pre-training method that can learn visual features from user click histories. Specifically, we propose a user interest reconstruction module to mine visual features related to user interests from behavior histories. We further propose a contrastive training method to avoid collapsing of embedding vectors. We conduct extensive experiments to verify that our method can learn users' visual interests, and our method achieves $0.46\%$ improvement in offline AUC and $0.88\%$ improvement in Taobao online GMV with p-value$<0.01$.

Abstract: Whole slide image (WSI) refers to a type of high-resolution scanned tissue image, which is extensively employed in computer-assisted diagnosis (CAD). The extremely high resolution and limited availability of region-level annotations make it challenging to employ deep learning methods for WSI-based digital diagnosis. Multiple instance learning (MIL) is a powerful tool to address the weak annotation problem, while Transformer has shown great success in the field of visual tasks. The combination of both should provide new insights for deep learning based image diagnosis. However, due to the limitations of single-level MIL and the attention mechanism's constraints on sequence length, directly applying Transformer to WSI-based MIL tasks is not practical. To tackle this issue, we propose a Multi-level MIL with Transformer (MMIL-Transformer) approach. By introducing a hierarchical structure to MIL, this approach enables efficient handling of MIL tasks that involve a large number of instances. To validate its effectiveness, we conducted a set of experiments on WSIs classification task, where MMIL-Transformer demonstrate superior performance compared to existing state-of-the-art methods. Our proposed approach achieves test AUC 94.74% and test accuracy 93.41% on CAMELYON16 dataset, test AUC 99.04% and test accuracy 94.37% on TCGA-NSCLC dataset, respectively. All code and pre-trained models are available at: https://github.com/hustvl/MMIL-Transformer

Abstract: In many information processing systems, it may be desirable to ensure that any change of the input, whether by shifting or scaling, results in a corresponding change in the system response. While deep neural networks are gradually replacing all traditional automatic processing methods, they surprisingly do not guarantee such normalization-equivariance (scale + shift) property, which can be detrimental in many applications. To address this issue, we propose a methodology for adapting existing neural networks so that normalization-equivariance holds by design. Our main claim is that not only ordinary convolutional layers, but also all activation functions, including the ReLU (rectified linear unit), which are applied element-wise to the pre-activated neurons, should be completely removed from neural networks and replaced by better conditioned alternatives. To this end, we introduce affine-constrained convolutions and channel-wise sort pooling layers as surrogates and show that these two architectural modifications do preserve normalization-equivariance without loss of performance. Experimental results in image denoising show that normalization-equivariant neural networks, in addition to their better conditioning, also provide much better generalization across noise levels.

Abstract: Each year, wildfires destroy larger areas of Spain, threatening numerous ecosystems. Humans cause 90% of them (negligence or provoked) and the behaviour of individuals is unpredictable. However, atmospheric and environmental variables affect the spread of wildfires, and they can be analysed by using deep learning. In order to mitigate the damage of these events we proposed the novel Wildfire Assessment Model (WAM). Our aim is to anticipate the economic and ecological impact of a wildfire, assisting managers resource allocation and decision making for dangerous regions in Spain, Castilla y Le\'on and Andaluc\'ia. The WAM uses a residual-style convolutional network architecture to perform regression over atmospheric variables and the greenness index, computing necessary resources, the control and extinction time, and the expected burnt surface area. It is first pre-trained with self-supervision over 100,000 examples of unlabelled data with a masked patch prediction objective and fine-tuned using 311 samples of wildfires. The pretraining allows the model to understand situations, outclassing baselines with a 1,4%, 3,7% and 9% improvement estimating human, heavy and aerial resources; 21% and 10,2% in expected extinction and control time; and 18,8% in expected burnt area. Using the WAM we provide an example assessment map of Castilla y Le\'on, visualizing the expected resources over an entire region.

Abstract: Existing pruning methods utilize the importance of each weight based on specified criteria only when searching for a sparse structure but do not utilize it during training. In this work, we propose a novel approach - \textbf{M}agnitude \textbf{A}ttention-based Dynamic \textbf{P}runing (MAP) method, which applies the importance of weights throughout both the forward and backward paths to explore sparse model structures dynamically. Magnitude attention is defined based on the magnitude of weights as continuous real-valued numbers enabling a seamless transition from a redundant to an effective sparse network by promoting efficient exploration. Additionally, the attention mechanism ensures more effective updates for important layers within the sparse network. In later stages of training, our approach shifts from exploration to exploitation, exclusively updating the sparse model composed of crucial weights based on the explored structure, resulting in pruned models that not only achieve performance comparable to dense models but also outperform previous pruning methods on CIFAR-10/100 and ImageNet.

Abstract: Multi-task visual perception has a wide range of applications in scene understanding such as autonomous driving. In this work, we devise an efficient unified framework to solve multiple common perception tasks, including instance segmentation, semantic segmentation, monocular 3D detection, and depth estimation. Simply sharing the same visual feature representations for these tasks impairs the performance of tasks, while independent task-specific feature extractors lead to parameter redundancy and latency. Thus, we design two feature-merge branches to learn feature basis, which can be useful to, and thus shared by, multiple perception tasks. Then, each task takes the corresponding feature basis as the input of the prediction task head to fulfill a specific task. In particular, one feature merge branch is designed for instance-level recognition the other for dense predictions. To enhance inter-branch communication, the instance branch passes pixel-wise spatial information of each instance to the dense branch using efficient dynamic convolution weighting. Moreover, a simple but effective dynamic routing mechanism is proposed to isolate task-specific features and leverage common properties among tasks. Our proposed framework, termed D2BNet, demonstrates a unique approach to parameter-efficient predictions for multi-task perception. In addition, as tasks benefit from co-training with each other, our solution achieves on par results on partially labeled settings on nuScenes and outperforms previous works for 3D detection and depth estimation on the Cityscapes dataset with full supervision.

Abstract: This paper presents a review of techniques used for the detection and tracking of UAVs or drones. There are different techniques that depend on collecting measurements of the position, velocity, and image of the UAV and then using them in detection and tracking. Hybrid detection techniques are also presented. The paper is a quick reference for a wide spectrum of methods that are used in the drone detection process.

Abstract: The augmentation parameters matter to few-shot semantic segmentation since they directly affect the training outcome by feeding the networks with varying perturbated samples. However, searching optimal augmentation parameters for few-shot segmentation models without annotations is a challenge that current methods fail to address. In this paper, we first propose a framework to determine the ``optimal'' parameters without human annotations by solving a distribution-matching problem between the intra-instance and intra-class similarity distribution, with the intra-instance similarity describing the similarity between the original sample of a particular anatomy and its augmented ones and the intra-class similarity representing the similarity between the selected sample and the others in the same class. Extensive experiments demonstrate the superiority of our optimized augmentation in boosting few-shot segmentation models. We greatly improve the top competing method by 1.27\% and 1.11\% on Abd-MRI and Abd-CT datasets, respectively, and even achieve a significant improvement for SSL-ALP on the left kidney by 3.39\% on the Abd-CT dataset.

Abstract: This work considers supervised learning to count from images and their corresponding point annotations. Where density-based counting methods typically use the point annotations only to create Gaussian-density maps, which act as the supervision signal, the starting point of this work is that point annotations have counting potential beyond density map generation. We introduce two methods that repurpose the available point annotations to enhance counting performance. The first is a counting-specific augmentation that leverages point annotations to simulate occluded objects in both input and density images to enhance the network's robustness to occlusions. The second method, foreground distillation, generates foreground masks from the point annotations, from which we train an auxiliary network on images with blacked-out backgrounds. By doing so, it learns to extract foreground counting knowledge without interference from the background. These methods can be seamlessly integrated with existing counting advances and are adaptable to different loss functions. We demonstrate complementary effects of the approaches, allowing us to achieve robust counting results even in challenging scenarios such as background clutter, occlusion, and varying crowd densities. Our proposed approach achieves strong counting results on multiple datasets, including ShanghaiTech Part\_A and Part\_B, UCF\_QNRF, JHU-Crowd++, and NWPU-Crowd.

Abstract: Image anonymization is widely adapted in practice to comply with privacy regulations in many regions. However, anonymization often degrades the quality of the data, reducing its utility for computer vision development. In this paper, we investigate the impact of image anonymization for training computer vision models on key computer vision tasks (detection, instance segmentation, and pose estimation). Specifically, we benchmark the recognition drop on common detection datasets, where we evaluate both traditional and realistic anonymization for faces and full bodies. Our comprehensive experiments reflect that traditional image anonymization substantially impacts final model performance, particularly when anonymizing the full body. Furthermore, we find that realistic anonymization can mitigate this decrease in performance, where our experiments reflect a minimal performance drop for face anonymization. Our study demonstrates that realistic anonymization can enable privacy-preserving computer vision development with minimal performance degradation across a range of important computer vision benchmarks.

Abstract: We introduce Mesogeos, a large-scale multi-purpose dataset for wildfire modeling in the Mediterranean. Mesogeos integrates variables representing wildfire drivers (meteorology, vegetation, human activity) and historical records of wildfire ignitions and burned areas for 17 years (2006-2022). It is designed as a cloud-friendly spatio-temporal dataset, namely a datacube, harmonizing all variables in a grid of 1km x 1km x 1-day resolution. The datacube structure offers opportunities to assess machine learning (ML) usage in various wildfire modeling tasks. We extract two ML-ready datasets that establish distinct tracks to demonstrate this potential: (1) short-term wildfire danger forecasting and (2) final burned area estimation given the point of ignition. We define appropriate metrics and baselines to evaluate the performance of models in each track. By publishing the datacube, along with the code to create the ML datasets and models, we encourage the community to foster the implementation of additional tracks for mitigating the increasing threat of wildfires in the Mediterranean.

Abstract: Reconstructing category-specific objects from a single image is a challenging task that requires inferring the geometry and appearance of an object from a limited viewpoint. Existing methods typically rely on local feature retrieval based on re-projection with known camera intrinsic, which are slow and prone to distortion at viewpoints distant from the input image. In this paper, we present Variable Radiance Field (VRF), a novel framework that can efficiently reconstruct category-specific objects from a single image without known camera parameters. Our key contributions are: (1) We parameterize the geometry and appearance of the object using a multi-scale global feature extractor, which avoids frequent point-wise feature retrieval and camera dependency. We also propose a contrastive learning-based pretraining strategy to improve the feature extractor. (2) We reduce the geometric complexity of the object by learning a category template, and use hypernetworks to generate a small neural radiance field for fast and instance-specific rendering. (3) We align each training instance to the template space using a learned similarity transformation, which enables semantic-consistent learning across different objects. We evaluate our method on the CO3D dataset and show that it outperforms existing methods in terms of quality and speed. We also demonstrate its applicability to shape interpolation and object placement tasks.

Abstract: Egocentric action recognition is essential for healthcare and assistive technology that relies on egocentric cameras because it allows for the automatic and continuous monitoring of activities of daily living (ADLs) without requiring any conscious effort from the user. This study explores the feasibility of using 2D hand and object pose information for egocentric action recognition. While current literature focuses on 3D hand pose information, our work shows that using 2D skeleton data is a promising approach for hand-based action classification, might offer privacy enhancement, and could be less computationally demanding. The study uses a state-of-the-art transformer-based method to classify sequences and achieves validation results of 94%, outperforming other existing solutions. The accuracy of the test subset drops to 76%, indicating the need for further generalization improvement. This research highlights the potential of 2D hand and object pose information for action recognition tasks and offers a promising alternative to 3D-based methods.

Abstract: The remarkable capabilities of pretrained image diffusion models have been utilized not only for generating fixed-size images but also for creating panoramas. However, naive stitching of multiple images often results in visible seams. Recent techniques have attempted to address this issue by performing joint diffusions in multiple windows and averaging latent features in overlapping regions. However, these approaches, which focus on seamless montage generation, often yield incoherent outputs by blending different scenes within a single image. To overcome this limitation, we propose SyncDiffusion, a plug-and-play module that synchronizes multiple diffusions through gradient descent from a perceptual similarity loss. Specifically, we compute the gradient of the perceptual loss using the predicted denoised images at each denoising step, providing meaningful guidance for achieving coherent montages. Our experimental results demonstrate that our method produces significantly more coherent outputs compared to previous methods (66.35% vs. 33.65% in our user study) while still maintaining fidelity (as assessed by GIQA) and compatibility with the input prompt (as measured by CLIP score).

Abstract: Transfer-based attack adopts the adversarial examples generated on the surrogate model to attack various models, making it applicable in the physical world and attracting increasing interest. Recently, various adversarial attacks have emerged to boost adversarial transferability from different perspectives. In this work, inspired by the fact that flat local minima are correlated with good generalization, we assume and empirically validate that adversarial examples at a flat local region tend to have good transferability by introducing a penalized gradient norm to the original loss function. Since directly optimizing the gradient regularization norm is computationally expensive and intractable for generating adversarial examples, we propose an approximation optimization method to simplify the gradient update of the objective function. Specifically, we randomly sample an example and adopt the first-order gradient to approximate the second-order Hessian matrix, which makes computing more efficient by interpolating two Jacobian matrices. Meanwhile, in order to obtain a more stable gradient direction, we randomly sample multiple examples and average the gradients of these examples to reduce the variance due to random sampling during the iterative process. Extensive experimental results on the ImageNet-compatible dataset show that the proposed method can generate adversarial examples at flat local regions, and significantly improve the adversarial transferability on either normally trained models or adversarially trained models than the state-of-the-art attacks.

Abstract: Unsupervised person re-identification has achieved great success through the self-improvement of individual neural networks. However, limited by the lack of diversity of discriminant information, a single network has difficulty learning sufficient discrimination ability by itself under unsupervised conditions. To address this limit, we develop a population-based evolutionary gaming (PEG) framework in which a population of diverse neural networks is trained concurrently through selection, reproduction, mutation, and population mutual learning iteratively. Specifically, the selection of networks to preserve is modeled as a cooperative game and solved by the best-response dynamics, then the reproduction and mutation are implemented by cloning and fluctuating hyper-parameters of networks to learn more diversity, and population mutual learning improves the discrimination of networks by knowledge distillation from each other within the population. In addition, we propose a cross-reference scatter (CRS) to approximately evaluate re-ID models without labeled samples and adopt it as the criterion of network selection in PEG. CRS measures a model's performance by indirectly estimating the accuracy of its predicted pseudo-labels according to the cohesion and separation of the feature space. Extensive experiments demonstrate that (1) CRS approximately measures the performance of models without labeled samples; (2) and PEG produces new state-of-the-art accuracy for person re-identification, indicating the great potential of population-based network cooperative training for unsupervised learning.

Abstract: Exploring robust and efficient association methods has always been an important issue in multiple-object tracking (MOT). Although existing tracking methods have achieved impressive performance, congestion and frequent occlusions still pose challenging problems in multi-object tracking. We reveal that performing sparse decomposition on dense scenes is a crucial step to enhance the performance of associating occluded targets. To this end, we propose a pseudo-depth estimation method for obtaining the relative depth of targets from 2D images. Secondly, we design a depth cascading matching (DCM) algorithm, which can use the obtained depth information to convert a dense target set into multiple sparse target subsets and perform data association on these sparse target subsets in order from near to far. By integrating the pseudo-depth method and the DCM strategy into the data association process, we propose a new tracker, called SparseTrack. SparseTrack provides a new perspective for solving the challenging crowded scene MOT problem. Only using IoU matching, SparseTrack achieves comparable performance with the state-of-the-art (SOTA) methods on the MOT17 and MOT20 benchmarks. Code and models are publicly available at \url{https://github.com/hustvl/SparseTrack}.

Abstract: Considering the balance of performance and efficiency, sampled point and voxel methods are usually employed to down-sample dense events into sparse ones. After that, one popular way is to leverage a graph model which treats the sparse points/voxels as nodes and adopts graph neural networks (GNNs) to learn the representation for event data. Although good performance can be obtained, however, their results are still limited mainly due to two issues. (1) Existing event GNNs generally adopt the additional max (or mean) pooling layer to summarize all node embeddings into a single graph-level representation for the whole event data representation. However, this approach fails to capture the importance of graph nodes and also fails to be fully aware of the node representations. (2) Existing methods generally employ either a sparse point or voxel graph representation model which thus lacks consideration of the complementary between these two types of representation models. To address these issues, in this paper, we propose a novel dual point-voxel absorbing graph representation learning for event stream data representation. To be specific, given the input event stream, we first transform it into the sparse event cloud and voxel grids and build dual absorbing graph models for them respectively. Then, we design a novel absorbing graph convolutional network (AGCN) for our dual absorbing graph representation and learning. The key aspect of the proposed AGCN is its ability to effectively capture the importance of nodes and thus be fully aware of node representations in summarizing all node representations through the introduced absorbing nodes. Finally, the event representations of dual learning branches are concatenated together to extract the complementary information of two cues. The output is then fed into a linear layer for event data classification.

Abstract: Scene analysis is essential for enabling autonomous systems, such as mobile robots, to operate in real-world environments. However, obtaining a comprehensive understanding of the scene requires solving multiple tasks, such as panoptic segmentation, instance orientation estimation, and scene classification. Solving these tasks given limited computing and battery capabilities on mobile platforms is challenging. To address this challenge, we introduce an efficient multi-task scene analysis approach, called EMSAFormer, that uses an RGB-D Transformer-based encoder to simultaneously perform the aforementioned tasks. Our approach builds upon the previously published EMSANet. However, we show that the dual CNN-based encoder of EMSANet can be replaced with a single Transformer-based encoder. To achieve this, we investigate how information from both RGB and depth data can be effectively incorporated in a single encoder. To accelerate inference on robotic hardware, we provide a custom NVIDIA TensorRT extension enabling highly optimization for our EMSAFormer approach. Through extensive experiments on the commonly used indoor datasets NYUv2, SUNRGB-D, and ScanNet, we show that our approach achieves state-of-the-art performance while still enabling inference with up to 39.1 FPS on an NVIDIA Jetson AGX Orin 32 GB.

Abstract: With the rapid development of geometric deep learning techniques, many mesh-based convolutional operators have been proposed to bridge irregular mesh structures and popular backbone networks. In this paper, we show that while convolutions are helpful, a simple architecture based exclusively on multi-layer perceptrons (MLPs) is competent enough to deal with mesh classification and semantic segmentation. Our new network architecture, named Mesh-MLP, takes mesh vertices equipped with the heat kernel signature (HKS) and dihedral angles as the input, replaces the convolution module of a ResNet with Multi-layer Perceptron (MLP), and utilizes layer normalization (LN) to perform the normalization of the layers. The all-MLP architecture operates in an end-to-end fashion and does not include a pooling module. Extensive experimental results on the mesh classification/segmentation tasks validate the effectiveness of the all-MLP architecture.

Abstract: Deep learning-based medical image segmentation models suffer from performance degradation when deployed to a new healthcare center. To address this issue, unsupervised domain adaptation and multi-source domain generalization methods have been proposed, which, however, are less favorable for clinical practice due to the cost of acquiring target-domain data and the privacy concerns associated with redistributing the data from multiple source domains. In this paper, we propose a \textbf{C}hannel-level \textbf{C}ontrastive \textbf{S}ingle \textbf{D}omain \textbf{G}eneralization (\textbf{C$^2$SDG}) model for medical image segmentation. In C$^2$SDG, the shallower features of each image and its style-augmented counterpart are extracted and used for contrastive training, resulting in the disentangled style representations and structure representations. The segmentation is performed based solely on the structure representations. Our method is novel in the contrastive perspective that enables channel-wise feature disentanglement using a single source domain. We evaluated C$^2$SDG against six SDG methods on a multi-domain joint optic cup and optic disc segmentation benchmark. Our results suggest the effectiveness of each module in C$^2$SDG and also indicate that C$^2$SDG outperforms the baseline and all competing methods with a large margin. The code will be available at \url{https://github.com/ShishuaiHu/CCSDG}.

Abstract: Current robotic hand manipulation narrowly operates with objects in predictable positions in limited environments. Thus, when the location of the target object deviates severely from the expected location, a robot sometimes responds in an unexpected way, especially when it operates with a human. For safe robot operation, we propose the EXit-aware Object Tracker (EXOT) on a robot hand camera that recognizes an object's absence during manipulation. The robot decides whether to proceed by examining the tracker's bounding box output containing the target object. We adopt an out-of-distribution classifier for more accurate object recognition since trackers can mistrack a background as a target object. To the best of our knowledge, our method is the first approach of applying an out-of-distribution classification technique to a tracker output. We evaluate our method on the first-person video benchmark dataset, TREK-150, and on the custom dataset, RMOT-223, that we collect from the UR5e robot. Then we test our tracker on the UR5e robot in real-time with a conveyor-belt sushi task, to examine the tracker's ability to track target dishes and to determine the exit status. Our tracker shows 38% higher exit-aware performance than a baseline method. The dataset and the code will be released at https://github.com/hskAlena/EXOT.

Abstract: The advent of large pre-trained models has brought about a paradigm shift in both visual representation learning and natural language processing. However, clustering unlabeled images, as a fundamental and classic machine learning problem, still lacks effective solution, particularly for large-scale datasets. In this paper, we propose a novel image clustering pipeline that leverages the powerful feature representation of large pre-trained models such as CLIP and cluster images effectively and efficiently at scale. We show that the pre-trained features are significantly more structured by further optimizing the rate reduction objective. The resulting features may significantly improve the clustering accuracy, e.g., from 57\% to 66\% on ImageNet-1k. Furthermore, by leveraging CLIP's image-text binding, we show how the new clustering method leads to a simple yet effective self-labeling algorithm that successfully works on unlabeled large datasets such as MS-COCO and LAION-Aesthetics. We will release the code in https://github.com/LeslieTrue/CPP.

Abstract: The quality of three-dimensional reconstruction is a key factor affecting the effectiveness of its application in areas such as virtual reality (VR) and augmented reality (AR) technologies. Neural Radiance Fields (NeRF) can generate realistic images from any viewpoint. It simultaneously reconstructs the shape, lighting, and materials of objects, and without surface defects, which breaks down the barrier between virtuality and reality. The potential spatial correspondences displayed by NeRF between reconstructed scenes and real-world scenes offer a wide range of practical applications possibilities. Despite significant progress in 3D reconstruction since NeRF were introduced, there remains considerable room for exploration and experimentation. NeRF-based models are susceptible to interference issues caused by colored "fog" noise. Additionally, they frequently encounter instabilities and failures while attempting to reconstruct unbounded scenes. Moreover, the model takes a significant amount of time to converge, making it even more challenging to use in such scenarios. Our approach, coined Enhance-NeRF, which adopts joint color to balance low and high reflectivity objects display, utilizes a decoding architecture with prior knowledge to improve recognition, and employs multi-layer performance evaluation mechanisms to enhance learning capacity. It achieves reconstruction of outdoor scenes within one hour under single-card condition. Based on experimental results, Enhance-NeRF partially enhances fitness capability and provides some support to outdoor scene reconstruction. The Enhance-NeRF method can be used as a plug-and-play component, making it easy to integrate with other NeRF-based models. The code is available at: https://github.com/TANQIanQ/Enhance-NeRF

Abstract: In this work, we present a pipeline to reconstruct the 3D pose of a horse from 4 simultaneous surveillance camera recordings. Our environment poses interesting challenges to tackle, such as limited field view of the cameras and a relatively closed and small environment. The pipeline consists of training a 2D markerless pose estimation model to work on every viewpoint, then applying it to the videos and performing triangulation. We present numerical evaluation of the results (error analysis), as well as show the utility of the achieved poses in downstream tasks of selected behavioral predictions. Our analysis of the predictive model for equine behavior showed a bias towards pain-induced horses, which aligns with our understanding of how behavior varies across painful and healthy subjects.

Abstract: This paper introduces a real-time GeoAI workflow for large-scale image analysis and the segmentation of Arctic permafrost features at a fine-granularity. Very high-resolution (0.5m) commercial imagery is used in this analysis. To achieve real-time prediction, our workflow employs a lightweight, deep learning-based instance segmentation model, SparseInst, which introduces and uses Instance Activation Maps to accurately locate the position of objects within the image scene. Experimental results show that the model can achieve better accuracy of prediction at a much faster inference speed than the popular Mask-RCNN model.

Abstract: Almost all advanced face swapping approaches use reconstruction as the proxy task, i.e., supervision only exists when the target and source belong to the same person. Otherwise, lacking pixel-level supervision, these methods struggle for source identity preservation. This paper proposes to construct reliable supervision, dubbed cycle triplets, which serves as the image-level guidance when the source identity differs from the target one during training. Specifically, we use face reenactment and blending techniques to synthesize the swapped face from real images in advance, where the synthetic face preserves source identity and target attributes. However, there may be some artifacts in such a synthetic face. To avoid the potential artifacts and drive the distribution of the network output close to the natural one, we reversely take synthetic images as input while the real face as reliable supervision during the training stage of face swapping. Besides, we empirically find that the existing methods tend to lose lower-face details like face shape and mouth from the source. This paper additionally designs a FixerNet, providing discriminative embeddings of lower faces as an enhancement. Our face swapping framework, named ReliableSwap, can boost the performance of any existing face swapping network with negligible overhead. Extensive experiments demonstrate the efficacy of our ReliableSwap, especially in identity preservation. The project page is https://reliable-swap.github.io/.

Abstract: Text-to-image generative models have enabled high-resolution image synthesis across different domains, but require users to specify the content they wish to generate. In this paper, we consider the inverse problem -- given a collection of different images, can we discover the generative concepts that represent each image? We present an unsupervised approach to discover generative concepts from a collection of images, disentangling different art styles in paintings, objects, and lighting from kitchen scenes, and discovering image classes given ImageNet images. We show how such generative concepts can accurately represent the content of images, be recombined and composed to generate new artistic and hybrid images, and be further used as a representation for downstream classification tasks.

Abstract: The field of image blending has gained significant popularity in recent years due to its ability to create visually stunning content. The main objective of image blending is to merge an object from one image onto another seamlessly, with minor masking adjustments. With the recent development of SAM, which can detect and segment targets in images automatically. Our approach (1) combines semantic object detection and segmentation with corresponding mask generation to automatically fuse images and (2) introduces the use of PAN for further quality enhancement during the fusion process. Our approach surpasses many classical visual fusion models in various performance indicators such as PSNR, SSIM, and Realism. Notably, our process is highly efficient and speedy, making it widely applicable in industrial settings. This new process has the potential to revolutionize visual content creation and improve productivity across various industries.

Abstract: This paper introduces a new large-scale image restoration dataset, called HQ-50K, which contains 50,000 high-quality images with rich texture details and semantic diversity. We analyze existing image restoration datasets from five different perspectives, including data scale, resolution, compression rates, texture details, and semantic coverage. However, we find that all of these datasets are deficient in some aspects. In contrast, HQ-50K considers all of these five aspects during the data curation process and meets all requirements. We also present a new Degradation-Aware Mixture of Expert (DAMoE) model, which enables a single model to handle multiple corruption types and unknown levels. Our extensive experiments demonstrate that HQ-50K consistently improves the performance on various image restoration tasks, such as super-resolution, denoising, dejpeg, and deraining. Furthermore, our proposed DAMoE, trained on our \dataset, outperforms existing state-of-the-art unified models designed for multiple restoration tasks and levels. The dataset and code are available at \url{https://github.com/littleYaang/HQ-50K}.

Abstract: In this paper, we propose the Matting Anything Model (MAM), an efficient and versatile framework for estimating the alpha matte of any instance in an image with flexible and interactive visual or linguistic user prompt guidance. MAM offers several significant advantages over previous specialized image matting networks: (i) MAM is capable of dealing with various types of image matting, including semantic, instance, and referring image matting with only a single model; (ii) MAM leverages the feature maps from the Segment Anything Model (SAM) and adopts a lightweight Mask-to-Matte (M2M) module to predict the alpha matte through iterative refinement, which has only 2.7 million trainable parameters. (iii) By incorporating SAM, MAM simplifies the user intervention required for the interactive use of image matting from the trimap to the box, point, or text prompt. We evaluate the performance of MAM on various image matting benchmarks, and the experimental results demonstrate that MAM achieves comparable performance to the state-of-the-art specialized image matting models under different metrics on each benchmark. Overall, MAM shows superior generalization ability and can effectively handle various image matting tasks with fewer parameters, making it a practical solution for unified image matting. Our code and models are open-sourced at https://github.com/SHI-Labs/Matting-Anything.

Abstract: Semantic 2D maps are commonly used by humans and machines for navigation purposes, whether it's walking or driving. However, these maps have limitations: they lack detail, often contain inaccuracies, and are difficult to create and maintain, especially in an automated fashion. Can we use raw imagery to automatically create better maps that can be easily interpreted by both humans and machines? We introduce SNAP, a deep network that learns rich neural 2D maps from ground-level and overhead images. We train our model to align neural maps estimated from different inputs, supervised only with camera poses over tens of millions of StreetView images. SNAP can resolve the location of challenging image queries beyond the reach of traditional methods, outperforming the state of the art in localization by a large margin. Moreover, our neural maps encode not only geometry and appearance but also high-level semantics, discovered without explicit supervision. This enables effective pre-training for data-efficient semantic scene understanding, with the potential to unlock cost-efficient creation of more detailed maps.

Abstract: A critical obstacle preventing NeRF models from being deployed broadly in the wild is their reliance on accurate camera poses. Consequently, there is growing interest in extending NeRF models to jointly optimize camera poses and scene representation, which offers an alternative to off-the-shelf SfM pipelines which have well-understood failure modes. Existing approaches for unposed NeRF operate under limited assumptions, such as a prior pose distribution or coarse pose initialization, making them less effective in a general setting. In this work, we propose a novel approach, LU-NeRF, that jointly estimates camera poses and neural radiance fields with relaxed assumptions on pose configuration. Our approach operates in a local-to-global manner, where we first optimize over local subsets of the data, dubbed mini-scenes. LU-NeRF estimates local pose and geometry for this challenging few-shot task. The mini-scene poses are brought into a global reference frame through a robust pose synchronization step, where a final global optimization of pose and scene can be performed. We show our LU-NeRF pipeline outperforms prior attempts at unposed NeRF without making restrictive assumptions on the pose prior. This allows us to operate in the general SE(3) pose setting, unlike the baselines. Our results also indicate our model can be complementary to feature-based SfM pipelines as it compares favorably to COLMAP on low-texture and low-resolution images.

Abstract: Vision-specific concepts such as "region" have played a key role in extending general machine learning frameworks to tasks like object detection. Given the success of region-based detectors for supervised learning and the progress of intra-image methods for contrastive learning, we explore the use of regions for reconstructive pre-training. Starting from Masked Autoencoding (MAE) both as a baseline and an inspiration, we propose a parallel pre-text task tailored to address the one-to-many mapping between images and regions. Since such regions can be generated in an unsupervised way, our approach (R-MAE) inherits the wide applicability from MAE, while being more "region-aware". We conduct thorough analyses during the development of R-MAE, and converge on a variant that is both effective and efficient (1.3% overhead over MAE). Moreover, it shows consistent quantitative improvements when generalized to various pre-training data and downstream detection and segmentation benchmarks. Finally, we provide extensive qualitative visualizations to enhance the understanding of R-MAE's behaviour and potential. Code will be made available at https://github.com/facebookresearch/r-mae.

Abstract: DDIM inversion has revealed the remarkable potential of real image editing within diffusion-based methods. However, the accuracy of DDIM reconstruction degrades as larger classifier-free guidance (CFG) scales being used for enhanced editing. Null-text inversion (NTI) optimizes null embeddings to align the reconstruction and inversion trajectories with larger CFG scales, enabling real image editing with cross-attention control. Negative-prompt inversion (NPI) further offers a training-free closed-form solution of NTI. However, it may introduce artifacts and is still constrained by DDIM reconstruction quality. To overcome these limitations, we propose Proximal Negative-Prompt Inversion (ProxNPI), extending the concepts of NTI and NPI. We enhance NPI with a regularization term and reconstruction guidance, which reduces artifacts while capitalizing on its training-free nature. Our method provides an efficient and straightforward approach, effectively addressing real image editing tasks with minimal computational overhead.

Abstract: Data association is a knotty problem for 2D Multiple Object Tracking due to the object occlusion. However, in 3D space, data association is not so hard. Only with a 3D Kalman Filter, the online object tracker can associate the detections from LiDAR. In this paper, we rethink the data association in 2D MOT and utilize the 3D object representation to separate each object in the feature space. Unlike the existing depth-based MOT methods, the 3D object representation can be jointly learned with the object association module. Besides, the object's 3D representation is learned from the video and supervised by the 2D tracking labels without additional manual annotations from LiDAR or pretrained depth estimator. With 3D object representation learning from Pseudo 3D object labels in monocular videos, we propose a new 2D MOT paradigm, called P3DTrack. Extensive experiments show the effectiveness of our method. We achieve new state-of-the-art performance on the large-scale Waymo Open Dataset.

Abstract: Driving scene understanding task involves detecting static elements such as lanes, traffic signs, and traffic lights, and their relationships with each other. To facilitate the development of comprehensive scene understanding solutions using multiple camera views, a new dataset called Road Genome (OpenLane-V2) has been released. This dataset allows for the exploration of complex road connections and situations where lane markings may be absent. Instead of using traditional lane markings, the lanes in this dataset are represented by centerlines, which offer a more suitable representation of lanes and their connections. In this study, we have introduced a new approach called TopoMask for predicting centerlines in road topology. Unlike existing approaches in the literature that rely on keypoints or parametric methods, TopoMask utilizes an instance-mask based formulation with a transformer-based architecture and, in order to enrich the mask instances with flow information, a direction label representation is proposed. TopoMask have ranked 4th in the OpenLane-V2 Score (OLS) and ranked 2nd in the F1 score of centerline prediction in OpenLane Topology Challenge 2023. In comparison to the current state-of-the-art method, TopoNet, the proposed method has achieved similar performance in Frechet-based lane detection and outperformed TopoNet in Chamfer-based lane detection without utilizing its scene graph neural network.

Abstract: With the advent of the big model era, the demand for data has become more important. Especially in monocular 3D object detection, expensive manual annotations potentially limit further developments. Existing works have investigated weakly supervised algorithms with the help of LiDAR modality to generate 3D pseudo labels, which cannot be applied to ordinary videos. In this paper, we propose a novel paradigm, termed as BA$^2$-Det, leveraging the idea of global-to-local 3D reconstruction for 2D supervised monocular 3D object detection. Specifically, we recover 3D structures from monocular videos by scene-level global reconstruction with global bundle adjustment (BA) and obtain object clusters by the DoubleClustering algorithm. Learning from completely reconstructed objects in global BA, GBA-Learner predicts pseudo labels for occluded objects. Finally, we train an LBA-Learner with object-centric local BA to generalize the generated 3D pseudo labels to moving objects. Experiments on the large-scale Waymo Open Dataset show that the performance of BA$^2$-Det is on par with the fully-supervised BA-Det trained with 10% videos and even outperforms some pioneer fully-supervised methods. We also show the great potential of BA$^2$-Det for detecting open-set 3D objects in complex scenes. The code will be made available. Project page: https://ba2det.site .

Abstract: This paper aims to deal with the ignored real-world complexities in prior work on human motion forecasting, emphasizing the social properties of multi-person motion, the diversity of motion and social interactions, and the complexity of articulated motion. To this end, we introduce a novel task of stochastic multi-person 3D motion forecasting. We propose a dual-level generative modeling framework that separately models independent individual motion at the local level and social interactions at the global level. Notably, this dual-level modeling mechanism can be achieved within a shared generative model, through introducing learnable latent codes that represent intents of future motion and switching the codes' modes of operation at different levels. Our framework is general; we instantiate it with different generative models, including generative adversarial networks and diffusion models, and various multi-person forecasting models. Extensive experiments on CMU-Mocap, MuPoTS-3D, and SoMoF benchmarks show that our approach produces diverse and accurate multi-person predictions, significantly outperforming the state of the art.

Abstract: We present a new test-time optimization method for estimating dense and long-range motion from a video sequence. Prior optical flow or particle video tracking algorithms typically operate within limited temporal windows, struggling to track through occlusions and maintain global consistency of estimated motion trajectories. We propose a complete and globally consistent motion representation, dubbed OmniMotion, that allows for accurate, full-length motion estimation of every pixel in a video. OmniMotion represents a video using a quasi-3D canonical volume and performs pixel-wise tracking via bijections between local and canonical space. This representation allows us to ensure global consistency, track through occlusions, and model any combination of camera and object motion. Extensive evaluations on the TAP-Vid benchmark and real-world footage show that our approach outperforms prior state-of-the-art methods by a large margin both quantitatively and qualitatively. See our project page for more results: http://omnimotion.github.io/

Abstract: Image recognition and generation have long been developed independently of each other. With the recent trend towards general-purpose representation learning, the development of general representations for both recognition and generation tasks is also promoted. However, preliminary attempts mainly focus on generation performance, but are still inferior on recognition tasks. These methods are modeled in the vector-quantized (VQ) space, whereas leading recognition methods use pixels as inputs. Our key insights are twofold: (1) pixels as inputs are crucial for recognition tasks; (2) VQ tokens as reconstruction targets are beneficial for generation tasks. These observations motivate us to propose an Alternating Denoising Diffusion Process (ADDP) that integrates these two spaces within a single representation learning framework. In each denoising step, our method first decodes pixels from previous VQ tokens, then generates new VQ tokens from the decoded pixels. The diffusion process gradually masks out a portion of VQ tokens to construct the training samples. The learned representations can be used to generate diverse high-fidelity images and also demonstrate excellent transfer performance on recognition tasks. Extensive experiments show that our method achieves competitive performance on unconditional generation, ImageNet classification, COCO detection, and ADE20k segmentation. Importantly, our method represents the first successful development of general representations applicable to both generation and dense recognition tasks. Code shall be released.

Abstract: High-quality instructions and responses are essential for the zero-shot performance of large language models on interactive natural language tasks. For interactive vision-language tasks involving intricate visual scenes, a large quantity of diverse and creative instruction-response pairs should be imperative to tune vision-language models (VLMs). Nevertheless, the current availability of vision-language instruction-response pairs in terms of quantity, diversity, and creativity remains limited, posing challenges to the generalization of interactive VLMs. Here we present MultI-Modal In-Context Instruction Tuning (MIMIC-IT), a dataset comprising 2.8 million multimodal instruction-response pairs, with 2.2 million unique instructions derived from images and videos. Each pair is accompanied by multi-modal in-context information, forming conversational contexts aimed at empowering VLMs in perception, reasoning, and planning. The instruction-response collection process, dubbed as Syphus, is scaled using an automatic annotation pipeline that combines human expertise with GPT's capabilities. Using the MIMIC-IT dataset, we train a large VLM named Otter. Based on extensive evaluations conducted on vision-language benchmarks, it has been observed that Otter demonstrates remarkable proficiency in multi-modal perception, reasoning, and in-context learning. Human evaluation reveals it effectively aligns with the user's intentions. We release the MIMIC-IT dataset, instruction-response collection pipeline, benchmarks, and the Otter model.

Abstract: Conversation agents fueled by Large Language Models (LLMs) are providing a new way to interact with visual data. While there have been initial attempts for image-based conversation models, this work addresses the underexplored field of video-based conversation by introducing Video-ChatGPT. It is a multimodal model that merges a video-adapted visual encoder with a LLM. The model is capable of understanding and generating human-like conversations about videos. We introduce a new dataset of 100,000 video-instruction pairs used to train Video-ChatGPT acquired via manual and semi-automated pipeline that is easily scalable and robust to label noise. We also develop a quantiative evaluation framework for video-based dialogue models to objectively analyse the strengths and weaknesses of proposed models. Our code, models, instruction-sets and demo are released at https://github.com/mbzuai-oryx/Video-ChatGPT.

Abstract: Estimating the depth of objects from a single image is a valuable task for many vision, robotics, and graphics applications. However, current methods often fail to produce accurate depth for objects in diverse scenes. In this work, we propose a simple yet effective Background Prompting strategy that adapts the input object image with a learned background. We learn the background prompts only using small-scale synthetic object datasets. To infer object depth on a real image, we place the segmented object into the learned background prompt and run off-the-shelf depth networks. Background Prompting helps the depth networks focus on the foreground object, as they are made invariant to background variations. Moreover, Background Prompting minimizes the domain gap between synthetic and real object images, leading to better sim2real generalization than simple finetuning. Results on multiple synthetic and real datasets demonstrate consistent improvements in real object depths for a variety of existing depth networks. Code and optimized background prompts can be found at: https://mbaradad.github.io/depth_prompt.

Abstract: Driven by scalable diffusion models trained on large-scale paired text-image datasets, text-to-image synthesis methods have shown compelling results. However, these models still fail to precisely follow the text prompt when multiple objects, attributes, and spatial compositions are involved in the prompt. In this paper, we identify the potential reasons in both the cross-attention and self-attention layers of the diffusion model. We propose two novel losses to refocus the attention maps according to a given layout during the sampling process. We perform comprehensive experiments on the DrawBench and HRS benchmarks using layouts synthesized by Large Language Models, showing that our proposed losses can be integrated easily and effectively into existing text-to-image methods and consistently improve their alignment between the generated images and the text prompts.