In the domain of scene graph generation, modeling commonsense as a single-prototype representation has been typically employed to facilitate the recognition of infrequent predicates. However, a fundamental challenge lies in the large intra-class variations of the visual appearance of predicates, resulting in subclasses within a predicate class. Such a challenge typically leads to the problem of misclassifying diverse predicates due to the rough predicate space clustering. In this paper, inspired by cognitive science, we maintain multi-prototype representations for each predicate class, which can accurately find the multiple class centers of the predicate space. Technically, we propose a novel multi-prototype learning framework consisting of three main steps: prototype-predicate matching, prototype updating, and prototype space optimization. We first design a triple-level optimal transport to match each predicate feature within the same class to a specific prototype. In addition, the prototypes are updated using momentum updating to find the class centers according to the matching results. Finally, we enhance the inter-class separability of the prototype space through iterations of the inter-class separability loss and intra-class compactness loss. Extensive evaluations demonstrate that our approach significantly outperforms state-of-the-art methods on the Visual Genome dataset.
论文检索
输入标题、作者或关键词，从 2,866 篇学术成果中精准定位
Whole slide image (WSI) classification is a crucial component in automated pathology analysis. Due to the inherent challenges of high-resolution WSIs and the absence of patch-level labels, most of the proposed methods follow the multiple instance learning (MIL) formulation. While MIL has been equipped with excellent instance feature extractors and aggregators, it is prone to learn spurious associations that undermine the performance of the model. For example, relying solely on color features may lead to erroneous diagnoses due to spurious associations between the disease and the color of patches. To address this issue, we develop a causal MIL framework for WSI classification, effectively distinguishing between causal and spurious associations. Specifically, we use the expectation of the intervention P(Y | do(X)) for bag prediction rather than the traditional likelihood P(Y | X). By applying the front-door adjustment, the spurious association is effectively blocked, where the intervened mediator is aggregated from patch-level features. We evaluate our proposed method on two publicly available WSI datasets, Camelyon16 and TCGA-NSCLC. Our causal MIL framework shows outstanding performance and is plug-and-play, seamlessly integrating with various feature extractors and aggregators.
Building scalable vision-language models to learn from diverse, multimodal data remains an open challenge. In this paper, we introduce an Efficient Vision-languagE foundation model, namely EVE, which is one unified multimodal Transformer pre-trained solely by one unified pre-training task. Specifically, EVE encodes both vision and language within a shared Transformer network integrated with modality-aware sparse Mixture-of-Experts (MoE) modules, which capture modality-specific information by selectively switching to different experts. To unify pre-training tasks of vision and language, EVE performs masked signal modeling on image-text pairs to reconstruct masked signals, i.e., image pixels and text tokens, given visible signals. This simple yet effective pre-training objective accelerates training by 4x compared to the model pre-trained with Image-Text Contrastive and Image-Text Matching losses. Owing to the combination of the unified architecture and pre-training task, EVE is easy to scale up, enabling better downstream performance with fewer resources and faster training speed. Despite its simplicity, EVE achieves state-of-the-art performance on various vision-language downstream tasks, including visual question answering, visual reasoning, and image-text retrieval.
TagFog: Textual Anchor Guidance and Fake Outlier Generation for Visual Out-of-Distribution Detection
PDF ↗Out-of-distribution (OOD) detection is crucial in many real-world applications. However, intelligent models are often trained solely on in-distribution (ID) data, leading to overconfidence when misclassifying OOD data as ID classes. In this study, we propose a new learning framework which leverage simple Jigsaw-based fake OOD data and rich semantic embeddings (`anchors') from the ChatGPT description of ID knowledge to help guide the training of the image encoder. The learning framework can be flexibly combined with existing post-hoc approaches to OOD detection, and extensive empirical evaluations on multiple OOD detection benchmarks demonstrate that rich textual representation of ID knowledge and fake OOD knowledge can well help train a visual encoder for OOD detection. With the learning framework, new state-of-the-art performance was achieved on all the benchmarks. The code is available at https://github.com/Cverchen/TagFog.
3D scene stylization refers to transform the appearance of a 3D scene to match a given style image, ensuring that images rendered from different viewpoints exhibit the same style as the given style image, while maintaining the 3D consistency of the stylized scene. Several existing methods have obtained impressive results in stylizing 3D scenes. However, the mod- els proposed by these methods need to be re-trained when applied to a new scene. In other words, their models are cou- pled with a specific scene and cannot adapt to arbitrary other scenes. To address this issue, we propose a novel 3D scene stylization framework to transfer an arbitrary style to an ar- bitrary scene, without any style-related or scene-related re- training. Concretely, we first map the appearance of the 3D scene into a 2D style pattern space, which realizes complete disentanglement of the geometry and appearance of the 3D scene and makes our model be generalized to arbitrary 3D scenes. Then we stylize the appearance of the 3D scene in the 2D style pattern space via a prompt-based 2D stylization al- gorithm. Experimental results demonstrate that our proposed framework is superior to SOTA methods in both visual qual- ity and generalization.
Consistency and interpretability have long been the critical issues in MRI reconstruction. While interpretability has been dramatically improved with the employment of deep unfolding networks (DUNs), current methods still suffer from inconsistencies and generate inferior anatomical structure. Especially in multi-contrast scenes, different imaging protocols often exacerbate the concerned issue. In this paper, we propose a range-null decomposition-assisted DUN architecture to ensure consistency while still providing desirable interpretability. Given the input decomposed, we argue that the inconsistency could be analytically relieved by feeding solely the null-space component into proximal mapping, while leaving the range-space counterpart fixed. More importantly, a correlation decoupling scheme is further proposed to narrow the information gap for multi-contrast fusion, which dynamically borrows isotropic features from the opponent while maintaining the modality-specific ones. Specifically, the two features are attached to different frequencies and learned individually by the newly designed isotropy encoder and anisotropy encoder. The former strives for the contrast-shared information, while the latter serves to capture the contrast-specific features. The quantitative and qualitative results show that our proposal outperforms most cutting-edge methods by a large margin. Codes will be released on https://github.com/chenjiachengzzz/RNU.
Low-level vision plays a crucial role in a wide range of imaging quality and image recognition applications. However, the limited size, quality, and diversity of datasets often pose significant challenges for low-level tasks. Data augmentation is the most effective and practical way of sample expansion, but the commonly used augmentation methods in high-level tasks have limited improvement in the low-level due to the boundary effects or the non-realistic context information. In this paper, we propose the Cut-and-Swap Frequency Components (CutFreq) method for low-level vision, which aims to preserve high-level representations with directionality and improve image synthesis quality. Observing the significant frequency domain differences between reconstructed images and real ones, in CutFreq, we propose to transform the input and real images separately in the frequency domain, then define two stages for the model training process, and finally swap the specified frequency bands respectively and inversely transform to generate augmented samples. The experimental results show the superior performance of CutFreq on five low-level vision tasks. Moreover, we demonstrate the effectiveness of CutFreq in the low-data regime. Code is available at https://github.com/DreamerCCC/CutFreq.
Consumer-grade cameras capture the RAW physical description of a scene and then process the image signals to obtain high-quality RGB images that are faithful to human visual perception. Conventionally, dense prediction scenes require high-precision recognition of objects in RGB images. However, predicting RGB data to exhibit the expected adaptability and robustness in harsh environments can be challenging. By capitalizing on the broader color gamut and higher bit depth offered by RAW data, in this paper, we demonstrate that RAW data can significantly improve the accuracy and robustness of object detectors in harsh environments. Firstly, we propose a general Pipeline for RAW Detection (PRD), along with a preprocessing strategy tailored to RAW data. Secondly, we design the RAW Corruption Benchmark (RCB) to address the dearth of benchmarks that reflect realistic scenarios in harsh environments. Thirdly, we demonstrate the significant improvement of RAW images in object detection for low-light and corrupt scenes. Specifically, our experiments indicate that PRD (using FCOS) outperforms RGB detection by 13.9mAP on LOD-Snow without generating restored images. Finally, we introduce a new nonlinear method called Functional Regularization (FR), which can effectively mine the unique characteristics of RAW data. The code is available at https://github.com/DreamerCCC/RawMining.
Group re-identification (G-ReID) aims to correctly associate groups with the same members captured by different cameras. However, supervised approaches for this task often suffer from the high cost of cross-camera sample labeling. Unsupervised methods based on clustering can avoid sample labeling, but the problem of member variations often makes clustering unstable, leading to incorrect pseudo-labels. To address these challenges, we propose an adaptive clustering-driven progressive learning approach (ACPL), which consists of a group adaptive clustering (GAC) module and a global dynamic prototype update (GDPU) module. Specifically, GAC designs the quasi-distance between groups, thus fully capitalizing on both individual-level and holistic information within groups. In the case of great uncertainty in intra-group members, GAC effectively minimizes the impact of non-discriminative features and reduces the noise in the model's pseudo-labels. Additionally, our GDPU devises a dynamic weight to update the prototypes and effectively mine the hard samples with complex member variations, which improves the model's robustness. Extensive experiments conducted on four popular G-ReID datasets demonstrate that our method not only achieves state-of-the-art performance on unsupervised G-ReID but also performs comparably to several fully supervised approaches.
Existing Transformer-based image deraining methods depend mostly on fixed single-input single-output U-Net architecture. In fact, this not only neglects the potentially explicit information from multiple image scales, but also lacks the capability of exploring the complementary implicit information across different scales. In this work, we rethink the multi-scale representations and design an effective multi-input multi-output framework that constructs intra- and inter-scale hierarchical modulation to better facilitate rain removal and help image restoration. We observe that rain levels reduce dramatically in coarser image scales, thus proposing to restore rain-free results from the coarsest scale to the finest scale in image pyramid inputs, which also alleviates the difficulty of model learning. Specifically, we integrate a sparsity-compensated Transformer block and a frequency-enhanced convolutional block into a coupled representation module, in order to jointly learn the intra-scale content-aware features. To facilitate representations learned at different scales to communicate with each other, we leverage a gated fusion module to adaptively aggregate the inter-scale spatial-aware features, which are rich in correlated information of rain appearances, leading to high-quality results. Extensive experiments demonstrate that our model achieves consistent gains on five benchmarks.
Recently, masked image modeling (MIM) has demonstrated promising prospects in self-supervised representation learning. However, existing MIM frameworks recover all masked patches equivalently, ignoring that the reconstruction difficulty of different patches can vary sharply due to their diverse distance from visible patches. In this paper, we propose a novel deep dynamic supervision to enable MIM methods to dynamically reconstruct patches with different degrees of difficulty at different pretraining phases and depths of the model. Our deep dynamic supervision helps to provide more locality inductive bias for ViTs especially in deep layers, which inherently makes up for the absence of local prior for self-attention mechanism. Built upon the deep dynamic supervision, we propose Deep Dynamic AutoEncoder (DDAE), a simple yet effective MIM framework that utilizes dynamic mechanisms for pixel regression and feature self-distillation simultaneously. Extensive experiments across a variety of vision tasks including ImageNet classification, semantic segmentation on ADE20K and object detection on COCO demonstrate the effectiveness of our approach.
Despite significant progress in utilizing pre-trained text-to-image diffusion models to guide the creation of 3D scenes, these methods often struggle to generate scenes that are sufficiently realistic, leading to "neural scene degeneration". In this work, we propose a new 3D scene generation model called Real3D. Specifically, Real3D designs a pipeline from a NeRF-like implicit renderer to a tetrahedrons-based explicit renderer, greatly improving the neural network's ability to generate various neural scenes. Moreover, Real3D introduces an additional discriminator to prevent neural scenes from falling into undesirable local optima, thus avoiding the degeneration phenomenon. Our experimental results demonstrate that Real3D outperforms all existing state-of-the-art text-to-3D generation methods, providing valuable insights to facilitate the development of learning-based 3D scene generation approaches.
Few-shot learning is a challenging task due to the limited availability of training samples. Recent few-shot learning studies with meta-learning and simple transfer learning methods have achieved promising performance. However, the feature extractor pre-trained with the upstream dataset may neglect the extraction of certain features which could be crucial for downstream tasks. In this study, inspired by the process of human learning in few-shot tasks, where humans not only observe the whole image (`global view') but also attend to various local image regions (`local view') for comprehensive understanding of detailed features, we propose a simple yet effective few-shot learning method called FeatWalk which can utilize the complementary nature of global and local views, therefore providing an intuitive and effective solution to the problem of insufficient local information extraction from the pre-trained feature extractor. Our method can be easily and flexibly combined with various existing methods, further enhancing few-shot learning performance. Extensive experiments on multiple benchmark datasets consistently demonstrate the effectiveness and versatility of our method.The source code is available at https://github.com/exceefind/FeatWalk.
Real-world image super-resolution (RWSR) is a long-standing problem as low-quality (LQ) images often have complex and unidentified degradations. Existing methods such as Generative Adversarial Networks (GANs) or continuous diffusion models present their own issues including GANs being difficult to train while continuous diffusion models requiring numerous inference steps. In this paper, we propose an Iterative Token Evaluation and Refinement (ITER) framework for RWSR, which utilizes a discrete diffusion model operating in the discrete token representation space, i.e., indexes of features extracted from a VQGAN codebook pre-trained with high-quality (HQ) images. We show that ITER is easier to train than GANs and more efficient than continuous diffusion models. Specifically, we divide RWSR into two sub-tasks, i.e., distortion removal and texture generation. Distortion removal involves simple HQ token prediction with LQ images, while texture generation uses a discrete diffusion model to iteratively refine the distortion removal output with a token refinement network. In particular, we propose to include a token evaluation network in the discrete diffusion process. It learns to evaluate which tokens are good restorations and helps to improve the iterative refinement results. Moreover, the evaluation network can first check status of the distortion removal output and then adaptively select total refinement steps needed, thereby maintaining a good balance between distortion removal and texture generation. Extensive experimental results show that ITER is easy to train and performs well within just 8 iterative steps.
Lane detection is to determine the precise location and shape of lanes on the road. Despite efforts made by current methods, it remains a challenging task due to the complexity of real-world scenarios. Existing approaches, whether proposal-based or keypoint-based, suffer from depicting lanes effectively and efficiently. Proposal-based methods detect lanes by distinguishing and regressing a collection of proposals in a streamlined top-down way, yet lack sufficient flexibility in lane representation. Keypoint-based methods, on the other hand, construct lanes flexibly from local descriptors, which typically entail complicated post-processing. In this paper, we present a “Sketch-and-Refine” paradigm that utilizes the merits of both keypoint-based and proposal-based methods. The motivation is that local directions of lanes are semantically simple and clear. At the “Sketch” stage, local directions of keypoints can be easily estimated by fast convolutional layers. Then we can build a set of lane proposals accordingly with moderate accuracy. At the “Refine” stage, we further optimize these proposals via a novel Lane Segment Association Module (LSAM), which allows adaptive lane segment adjustment. Last but not least, we propose multi-level feature integration to enrich lane feature representations more efficiently. Based on the proposed “Sketch-and-Refine” paradigm, we propose a fast yet effective lane detector dubbed “SRLane”. Experiments show that our SRLane can run at a fast speed (i.e., 278 FPS) while yielding an F1 score of 78.9%. The source code is available at: https://github.com/passerer/SRLane.
Improving Robustness for Joint Optimization of Camera Pose and Decomposed Low-Rank Tensorial Radiance Fields
PDF ↗In this paper, we propose an algorithm that allows joint refinement of camera pose and scene geometry represented by decomposed low-rank tensor, using only 2D images as supervision. First, we conduct a pilot study based on a 1D signal and relate our findings to 3D scenarios, where the naive joint pose optimization on voxel-based NeRFs can easily lead to sub-optimal solutions. Moreover, based on the analysis of the frequency spectrum, we propose to apply convolutional Gaussian filters on 2D and 3D radiance fields for a coarse-to-fine training schedule that enables joint camera pose optimization. Leveraging the decomposition property in decomposed low-rank tensor, our method achieves an equivalent effect to brute-force 3D convolution with only incurring little computational overhead. To further improve the robustness and stability of joint optimization, we also propose techniques of smoothed 2D supervision, randomly scaled kernel parameters, and edge-guided loss mask. Extensive quantitative and qualitative evaluations demonstrate that our proposed framework achieves superior performance in novel view synthesis as well as rapid convergence for optimization. The source code is available at https://github.com/Nemo1999/Joint-TensoRF.
Lane detection is a critical task in autonomous driving, which requires accurately predicting the complex topology of lanes in various scenarios. While previous methods of lane detection have shown success, challenges still exist, especially in scenarios where lane markings are absent. In this paper, we analyze the role of global and local features in accurately detecting lanes and propose a Hybrid Global-Local Perception Network (HGLNet) to leverage them. Global and local features play distinct roles in lane detection by respectively aiding in the detection of lane instances and the localization of corresponding lanes. HGLNet extracts global semantic context by utilizing a global extraction head that aggregates information about adaptive sampling points around lanes, achieving an optimal trade-off between performance and efficiency. Moreover, we introduce a Multi-hierarchy feature aggregator (MFA) to capture feature hierarchies in both regional and local ranges, elevating the representation of local features. The proposed Hybrid architecture can simultaneously focus on global and local features at different depth levels and efficiently integrate them to sense the global presence of lanes and accurately regress their locations. Experimental results demonstrate that our proposed method improves detection accuracy in various challenging scenarios, outperforming the state-of-the-art lane detection methods.
Recent LiDAR-based 3D Object Detection (3DOD) methods show promising results, but they often do not generalize well to target domains outside the source (or training) data distribution. To reduce such domain gaps and thus to make 3DOD models more generalizable, we introduce a novel unsupervised domain adaptation (UDA) method, called CMDA, which (i) leverages visual semantic cues from an image modality (i.e., camera images) as an effective semantic bridge to close the domain gap in the cross-modal Bird's Eye View (BEV) representations. Further, (ii) we also introduce a self-training-based learning strategy, wherein a model is adversarially trained to generate domain-invariant features, which disrupt the discrimination of whether a feature instance comes from a source or an unseen target domain. Overall, our CMDA framework guides the 3DOD model to generate highly informative and domain-adaptive features for novel data distributions. In our extensive experiments with large-scale benchmarks, such as nuScenes, Waymo, and KITTI, those mentioned above provide significant performance gains for UDA tasks, achieving state-of-the-art performance.
Fine Structure-Aware Sampling: A New Sampling Training Scheme for Pixel-Aligned Implicit Models in Single-View Human Reconstruction
PDF ↗Pixel-aligned implicit models, such as PIFu, PIFuHD, and ICON, are used for single-view clothed human reconstruction. These models need to be trained using a sampling training scheme. Existing sampling training schemes either fail to capture thin surfaces (e.g. ears, fingers) or cause noisy artefacts in reconstructed meshes. To address these problems, we introduce Fine Structured-Aware Sampling (FSS), a new sampling training scheme to train pixel-aligned implicit models for single-view human reconstruction. FSS resolves the aforementioned problems by proactively adapting to the thickness and complexity of surfaces. In addition, unlike existing sampling training schemes, FSS shows how normals of sample points can be capitalized in the training process to improve results. Lastly, to further improve the training process, FSS proposes a mesh thickness loss signal for pixel-aligned implicit models. It becomes computationally feasible to introduce this loss once a slight reworking of the pixel-aligned implicit function framework is carried out. Our results show that our methods significantly outperform SOTA methods qualitatively and quantitatively. Our code is publicly available at https://github.com/kcyt/FSS.
A significant volume of analog information, i.e., documents and images, have been digitized in the form of scanned copies for storing, sharing, and/or analyzing in the digital world. However, the quality of such contents is severely degraded by various distortions caused by printing, storing, and scanning processes in the physical world. Although restoring high-quality content from scanned copies has become an indispensable task for many products, it has not been systematically explored, and to the best of our knowledge, no public datasets are available. In this paper, we define this problem as Descanning and introduce a new high-quality and large-scale dataset named DESCAN-18K. It contains 18K pairs of original and scanned images collected in the wild containing multiple complex degradations. In order to eliminate such complex degradations, we propose a new image restoration model called DescanDiffusion consisting of a color encoder that corrects the global color degradation and a conditional denoising diffusion probabilistic model (DDPM) that removes local degradations. To further improve the generalization ability of DescanDiffusion, we also design a synthetic data generation scheme by reproducing prominent degradations in scanned images. We demonstrate that our DescanDiffusion outperforms other baselines including commercial restoration products, objectively and subjectively, via comprehensive experiments and analyses.