CVPR 2024 Papers

High-resolution wide-angle fisheye images are becoming more and more important for robotics applications such as autonomous driving. However using ordinary convolutional neural networks or vision transformers on this data is problematic due to projection and distortion losses introduced when projecting to a rectangular grid on the plane. We introduce the HEAL-SWIN transformer which combines the highly uniform Hierarchical Equal Area iso-Latitude Pixelation (HEALPix) grid used in astrophysics and cosmology with the Hierarchical Shifted-Window (SWIN) transformer to yield an efficient and flexible model capable of training on high-resolution distortion-free spherical data. In HEAL-SWIN the nested structure of the HEALPix grid is used to perform the patching and windowing operations of the SWIN transformer enabling the network to process spherical representations with minimal computational overhead. We demonstrate the superior performance of our model on both synthetic and real automotive datasets as well as a selection of other image datasets for semantic segmentation depth regression and classification tasks. Our code is publicly available.

Vision-language foundation models have shown remarkable performance in various zero-shot settings such as image retrieval classification or captioning. But so far those models seem to fall behind when it comes to zero-shot localization of referential expressions and objects in images. As a result they need to be fine-tuned for this task. In this paper we show that pretrained vision-language (VL) models allow for zero-shot open-vocabulary object localization without any fine-tuning. To leverage those capabilities we propose a Grounding Everything Module (GEM) that generalizes the idea of value-value attention introduced by CLIPSurgery to a self-self attention path. We show that the concept of self-self attention corresponds to clustering thus enforcing groups of tokens arising from the same object to be similar while preserving the alignment with the language space. To further guide the group formation we propose a set of regularizations that allows the model to finally generalize across datasets and backbones. We evaluate the proposed GEM framework on various benchmark tasks and datasets for semantic segmentation. GEM not only outperforms other training-free open-vocabulary localization methods but also achieves state-of-the-art results on the recently proposed OpenImagesV7 large-scale segmentation benchmark. Code is available at https://github.com/WalBouss/GEM

Vision graph neural networks (ViG) offer a new avenue for exploration in computer vision. A major bottleneck in ViGs is the inefficient k-nearest neighbor (KNN) operation used for graph construction. To solve this issue we propose a new method for designing ViGs Dynamic Axial Graph Construction (DAGC) which is more efficient than KNN as it limits the number of considered graph connections made within an image. Additionally we propose a novel CNN-GNN architecture GreedyViG which uses DAGC. Extensive experiments show that GreedyViG beats existing ViG CNN and ViT architectures in terms of accuracy GMACs and parameters on image classification object detection instance segmentation and semantic segmentation tasks. Our smallest model GreedyViG-S achieves 81.1% top-1 accuracy on ImageNet-1K 2.9% higher than Vision GNN and 2.2% higher than Vision HyperGraph Neural Network (ViHGNN) with less GMACs and a similar number of parameters. Our largest model GreedyViG-B obtains 83.9% top-1 accuracy 0.2% higher than Vision GNN with a 66.6% decrease in parameters and a 69% decrease in GMACs. GreedyViG-B also obtains the same accuracy as ViHGNN with a 67.3% decrease in parameters and a 71.3% decrease in GMACs. Our work shows that hybrid CNN-GNN architectures not only provide a new avenue for designing efficient models but that they can also exceed the performance of current state-of-the-art models.

Recent advancements in Large Vision-Language Models (VLMs) have shown great promise in natural image domains allowing users to hold a dialogue about given visual content. However such general-domain VLMs perform poorly for Remote Sensing (RS) scenarios leading to inaccurate or fabricated information when presented with RS domain-specific queries. Such a behavior emerges due to the unique challenges introduced by RS imagery. For example to handle high-resolution RS imagery with diverse scale changes across categories and many small objects region-level reasoning is necessary alongside holistic scene interpretation. Furthermore the lack of domain-specific multimodal instruction following data as well as strong backbone models for RS make it hard for the models to align their behavior with user queries. To address these limitations we propose GeoChat - the first versatile remote sensing VLM that offers multitask conversational capabilities with high-resolution RS images. Specifically GeoChat can not only answer image-level queries but also accepts region inputs to hold region-specific dialogue. Furthermore it can visually ground objects in its responses by referring to their spatial coordinates. To address the lack of domain-specific datasets we generate a novel RS multimodal instruction-following dataset by extending image-text pairs from existing diverse RS datasets. Leveraging this rich dataset we fine-tune our remote sensing VLM based on the LLaVA-1.5 architecture. We establish a comprehensive benchmark for RS multitask conversations and compare with a number of baseline methods. GeoChat demonstrates robust zero-shot performance on various remote sensing tasks e.g. image and region captioning visual question answering scene classification visually grounded conversations and referring object detection. Our codes will be open-sourced.

This paper proposes a GeneraLIst encoder-Decoder (GLID) pre-training method for better handling various downstream computer vision tasks. While self-supervised pre-training approaches e.g. Masked Autoencoder have shown success in transfer learning task-specific sub-architectures are still required to be appended for different downstream tasks which cannot enjoy the benefits of large-scale pre-training. GLID overcomes this challenge by allowing the pre-trained generalist encoder-decoder to be fine-tuned on various vision tasks with minimal task-specific architecture modifications. In the GLID training scheme pre-training pretext task and other downstream tasks are modeled as "query-to-answer" problems including the pre-training pretext task and other downstream tasks. We pre-train a task-agnostic encoder-decoder with query-mask pairs. During fine-tuning GLID maintains the pre-trained encoder-decoder and queries only replacing the topmost linear transformation layer with task-specific linear heads. This minimizes the pretrain-finetune architecture inconsistency and enables the pre-trained model to better adapt to downstream tasks. GLID achieves competitive performance on various vision tasks including object detection image segmentation pose estimation and depth estimation outperforming or matching specialist models such as Mask2Former DETR ViTPose and BinsFormer.

Principal component analysis (PCA) along with its extensions to manifolds and outlier contaminated data have been indispensable in computer vision and machine learning. In this work we present a unifying formalism for PCA and its variants and introduce a framework based on the flags of linear subspaces i.e. a hierarchy of nested linear subspaces of increasing dimension which not only allows for a common implementation but also yields novel variants not explored previously. We begin by generalizing traditional PCA methods that either maximize variance or minimize reconstruction error. We expand these interpretations to develop a wide array of new dimensionality reduction algorithms by accounting for outliers and the data manifold. To devise a common computational approach we recast robust and dual forms of PCA as optimization problems on flag manifolds. We then integrate tangent space approximations of principal geodesic analysis (tangent-PCA) into this flag-based framework creating novel robust and dual geodesic PCA variations. The remarkable flexibility offered by the `flagification' introduced here enables even more algorithmic variants identified by specific flag types. Last but not least we propose an effective convergent solver for these flag-formulations employing the Stiefel manifold. Our empirical results on both real-world and synthetic scenarios demonstrate the superiority of our novel algorithms especially in terms of robustness to outliers on manifolds.

Polarization is a fundamental property of light that encodes abundant information regarding surface shape material illumination and viewing geometry. The computer vision community has witnessed a blossom of polarization-based vision applications such as reflection removal shape-from-polarization (SfP) transparent object segmentation and color constancy partially due to the emergence of single-chip mono/color polarization sensors that make polarization data acquisition easier than ever. However is polarization-based vision vulnerable to adversarial attacks? If so is that possible to realize these adversarial attacks in the physical world without being perceived by human eyes? In this paper we warn the community of the vulnerability of polarization-based vision which can be more serious than RGB-based vision. By adapting a commercial LCD projector we achieve locally controllable polarizing projection which is successfully utilized to fool state-of-the-art polarization-based vision algorithms for glass segmentation and SfP. Compared with existing physical attacks on RGB-based vision which always suffer from the trade-off between attack efficacy and eye conceivability the adversarial attackers based on polarizing projection are contact-free and visually imperceptible since naked human eyes can rarely perceive the difference of viciously manipulated polarizing light and ordinary illumination. This poses unprecedented risks on polarization-based vision for which due attentions should be paid and counter measures be considered.

We introduce Florence-2 a novel vision foundation model with a unified prompt-based representation for various computer vision and vision-language tasks. While existing large vision models excel in transfer learning they struggle to perform diverse tasks with simple instructions a capability that implies handling the complexity of various spatial hierarchy and semantic granularity. Florence-2 was designed to take text-prompt as task instructions and generate desirable results in text forms whether it be captioning object detection grounding or segmentation. This multi-task learning setup demands large-scale high-quality annotated data. To this end we co-developed FLD-5B that consists of 5.4 billion comprehensive visual annotations on 126 million images using an iterative strategy of automated image annotation and model refinement. We adopted a sequence-to-sequence structure to train Florence-2 to perform versatile and comprehensive vision tasks. Extensive evaluations on numerous tasks demonstrated Florence-2 to be a strong vision foundation model contender with unprecedented zero-shot and fine-tuning capabilities.

Affine subspaces of Euclidean spaces are also referred to as flats. A standard task in computer vision or more generally in engineering and applied sciences is fitting a flat to a set of points which is commonly solved using the PCA. We generalize this technique to enable fitting a flat to a set of other flats possibly of varying dimensions based on representing the flats as squared distance fields. Compared to previous approaches such as Riemannian centers of mass in the manifold of affine Grassmannians our approach is conceptually much simpler and computationally more efficient yet offers desirable properties such as respecting symmetries and being equivariant to rigid transformations leading to more intuitive and useful results in practice. We demonstrate these claims in a number of synthetic experiments and a multi-view reconstruction task of line-like objects.

The Laplace-Beltrami operator (LBO) emerges from studying manifolds equipped with a Riemannian metric. It is often called the swiss army knife of geometry processing as it allows to capture intrinsic shape information and gives rise to heat diffusion geodesic distances and a multitude of shape descriptors. It also plays a central role in geometric deep learning. In this work we explore Finsler manifolds as a generalization of Riemannian manifolds. We revisit the Finsler heat equation and derive a Finsler heat kernel and a Finsler-Laplace-Beltrami Operator (FLBO): a novel theoretically justified anisotropic Laplace-Beltrami operator (ALBO). In experimental evaluations we demonstrate that the proposed FLBO is a valuable alternative to the traditional Riemannian-based LBO and ALBOs for spatial filtering and shape correspondence estimation. We hope that the proposed Finsler heat kernel and the FLBO will inspire further exploration of Finsler geometry in the computer vision community.

Event sensors offer high temporal resolution visual sensing which makes them ideal for perceiving fast visual phenomena without suffering from motion blur. Certain applications in robotics and vision-based navigation require 3D perception of an object undergoing circular or spinning motion in front of a static camera such as recovering the angular velocity and shape of the object. The setting is equivalent to observing a static object with an orbiting camera. In this paper we propose event-based structure-from-orbit (eSfO) where the aim is to simultaneously reconstruct the 3D structure of a fast spinning object observed from a static event camera and recover the equivalent orbital motion of the camera. Our contributions are threefold: since state-of-the-art event feature trackers cannot handle periodic self-occlusion due to the spinning motion we develop a novel event feature tracker based on spatio-temporal clustering and data association that can better track the helical trajectories of valid features in the event data. The feature tracks are then fed to our novel factor graph-based structure-from-orbit back-end that calculates the orbital motion parameters (e.g. spin rate relative rotational axis) that minimize the reprojection error. For evaluation we produce a new event dataset of objects under spinning motion. Comparisons against ground truth indicate the efficacy of eSfO.

We propose Strongly Supervised pre-training with ScreenShots (S4) - a novel pre-training paradigm for Vision-Language Models using data from large-scale web screenshot rendering. Using web screenshots unlocks a treasure trove of visual and textual cues that are not present in using image-text pairs. In S4 we leverage the inherent tree-structured hierarchy of HTML elements and the spatial localization to carefully design 10 pre-training tasks with large scale annotated data. These tasks resemble downstream tasks across different domains and the annotations are cheap to obtain. We demonstrate that compared to current screenshot pre-training objectives our innovative pre-training method significantly enhances performance of image-to-text model in nine varied and popular downstream tasks - up to 76.1% improvements on Table Detection and at least 1% on Widget Captioning.

Vision-language models (VLMs) have recently shown promising results in traditional downstream tasks. Evaluation studies have emerged to assess their abilities with the majority focusing on the third-person perspective and only a few addressing specific tasks from the first-person perspective. However the capability of VLMs to "think" from a first-person perspective a crucial attribute for advancing autonomous agents and robotics remains largely unexplored. To bridge this research gap we introduce EgoThink a novel visual question-answering benchmark that encompasses six core capabilities with twelve detailed dimensions. The benchmark is constructed using selected clips from egocentric videos with manually annotated question-answer pairs containing first-person information. To comprehensively assess VLMs we evaluate twenty-one popular VLMs on EgoThink. Moreover given the open-ended format of the answers we use GPT-4 as the automatic judge to compute single-answer grading. Experimental results indicate that although GPT-4V leads in numerous dimensions all evaluated VLMs still possess considerable potential for improvement in first-person perspective tasks. Meanwhile enlarging the number of trainable parameters has the most significant impact on model performance on EgoThink. In conclusion EgoThink serves as a valuable addition to existing evaluation benchmarks for VLMs providing an indispensable resource for future research in the realm of embodied artificial intelligence and robotics.

Multi-task learning for dense prediction has emerged as a pivotal area in computer vision enabling simultaneous processing of diverse yet interrelated pixel-wise prediction tasks. However the substantial computational demands of state-of-the-art (SoTA) models often limit their widespread deployment. This paper addresses this challenge by introducing network binarization to compress resource-intensive multi-task dense predictors. Specifically our goal is to significantly accelerate multi-task dense prediction models via Binary Neural Networks (BNNs) while maintaining and even improving model performance at the same time. To reach this goal we propose a Binary Multi-task Dense Predictor Bi-MTDP and several variants of \bimtdp in which a multi-task dense predictor is constructed via specified binarized modules. Our systematical analysis of this predictor reveals that performance drop from binarization is primarily caused by severe information degradation. To address this issue we introduce a deep information bottleneck layer that enforces representations for downstream tasks satisfying Gaussian distribution in forward propagation. Moreover we introduce a knowledge distillation mechanism to correct the direction of information flow in backward propagation. Intriguingly one variant of Bi-MTDP outperforms full-precision (FP) multi-task dense prediction SoTAs ARTC (CNN-based) and InvPT (ViT-based). This result indicates that Bi-MTDP is not merely a naive trade-off between performance and efficiency but is rather a benefit of the redundant information flow thanks to the multi-task architecture.

We introduce Deformable Convolution v4 (DCNv4) a highly efficient and effective operator designed for a broad spectrum of vision applications. DCNv4 addresses the limitations of its predecessor DCNv3 with two key enhancements: 1. removing softmax normalization in spatial aggregation to enhance its dynamic property and expressive power and 2. optimizing memory access to minimize redundant operations for speedup. These improvements result in a significantly faster convergence compared to DCNv3 and a substantial increase in processing speed with DCNv4 achieving more than three times the forward speed. DCNv4 demonstrates exceptional performance across various tasks including image classification instance and semantic segmentation and notably image generation. When integrated into generative models like U-Net in the latent diffusion model DCNv4 outperforms its baseline underscoring its possibility to enhance generative models. In practical applications replacing DCNv3 with DCNv4 in the InternImage model to create FlashInternImage results in up to 80% speed increase and further performance improvement without further modifications. The advancements in speed and efficiency of DCNv4 combined with its robust performance across diverse vision tasks show its potential as a foundational building block for future vision models.

The recent advance in vision-language models is largely attributed to the abundance of image-text data. We aim to replicate this success for video-language models but there simply is not enough human-curated video-text data available. We thus resort to fine-tuning a video-language model from a strong image-language baseline with synthesized instructional data. The resulting video model by video-instruction-tuning (VIIT) is then used to auto-label millions of videos to generate high-quality captions. We show the adapted video-language model performs well on a wide range of video-language benchmarks. For instance it surpasses the best prior result on open-ended NExT-QA by2.8%. Besides our model generates detailed descriptions for previously unseen videos which provide better textual supervision than existing methods. Experiments show that a video-language dual-encoder model contrastively trained on these auto-generated captions is 3.8% better than the strongest baseline that also leverages vision-language models. Our best model outperforms state-of-the-art methods on MSR-VTT zero-shot text-to-video retrieval by 6%. As a side product we generate the largest video caption dataset to date.

Vision Transformer (ViT) has emerged as a prominent architecture for various computer vision tasks. In ViT we divide the input image into patch tokens and process them through a stack of self-attention blocks. However unlike Convolutional Neural Network (CNN) ViT's simple architecture has no informative inductive bias (e.g. locality etc.). Due to this ViT requires a large amount of data for pre-training. Various data-efficient approaches (DeiT) have been proposed to train ViT on balanced datasets effectively. However limited literature discusses the use of ViT for datasets with long-tailed imbalances. In this work we introduce DeiT-LT to tackle the problem of training ViTs from scratch on long-tailed datasets. In DeiT-LT we introduce an efficient and effective way of distillation from CNN via distillation \texttt DIST token by using out-of-distribution images and re-weighting the distillation loss to enhance focus on tail classes. This leads to the learning of local CNN-like features in early ViT blocks improving generalization for tail classes. Further to mitigate overfitting we propose distilling from a flat CNN teacher which leads to learning low-rank generalizable features for DIST tokens across all ViT blocks. With the proposed DeiT-LT scheme the distillation DIST token becomes an expert on the tail classes and the classifier CLS token becomes an expert on the head classes. The experts help to effectively learn features corresponding to both the majority and minority classes using a distinct set of tokens within the same ViT architecture. We show the effectiveness of DeiT-LT for training ViT from scratch on datasets ranging from small-scale CIFAR-10 LT to large-scale iNaturalist-2018. Project Page: https://rangwani-harsh.github.io/DeiT-LT.

Understanding the dynamics of generic 3D scenes is fundamentally challenging in computer vision essential in enhancing applications related to scene reconstruction motion tracking and avatar creation. In this work we address the task as the problem of inferring dense long-range motion of 3D points. By observing a set of point trajectories we aim to learn an implicit motion field parameterized by a neural network to predict the movement of novel points within the same domain without relying on any data-driven or scene-specific priors. To achieve this our approach builds upon the recently introduced dynamic point field model that learns smooth deformation fields between the canonical frame and individual observation frames. However temporal consistency between consecutive frames is neglected and the number of required parameters increases linearly with the sequence length due to per-frame modeling. To address these shortcomings we exploit the intrinsic regularization provided by SIREN and modify the input layer to produce a spatiotemporally smooth motion field. Additionally we analyze the motion field Jacobian matrix and discover that the motion degrees of freedom (DOFs) in an infinitesimal area around a point and the network hidden variables have different behaviors to affect the model's representational power. This enables us to improve the model representation capability while retaining the model compactness. Furthermore to reduce the risk of overfitting we introduce a regularization term based on the assumption of piece-wise motion smoothness. Our experiments assess the model's performance in predicting unseen point trajectories and its application in temporal mesh alignment with guidance. The results demonstrate its superiority and effectiveness. The code and data for the project are publicly available at https://yz-cnsdqz.github.io/eigenmotion/DOMA.

Multi-view stereo reconstruction (MVS) in the wild requires to first estimate the camera intrinsic and extrinsic parameters. These are usually tedious and cumbersome to obtain yet they are mandatory to triangulate corresponding pixels in 3D space which is at the core of all best performing MVS algorithms. In this work we take an opposite stance and introduce DUSt3R a radically novel paradigm for Dense and Unconstrained Stereo 3D Reconstruction of arbitrary image collections operating without prior information about camera calibration nor viewpoint poses. We cast the pairwise reconstruction problem as a regression of pointmaps relaxing the hard constraints of usual projective camera models. We show that this formulation smoothly unifies the monocular and binocular reconstruction cases. In the case where more than two images are provided we further propose a simple yet effective global alignment strategy that expresses all pairwise pointmaps in a common reference frame. We base our network architecture on standard Transformer encoders and decoders allowing us to leverage powerful pretrained models. Our formulation directly provides a 3D model of the scene as well as depth information but interestingly we can seamlessly recover from it pixel matches focal lengths relative and absolute cameras. Extensive experiments on all these tasks showcase how DUSt3R effectively unifies various 3D vision tasks setting new performance records on monocular & multi-view depth estimation as well as relative pose estimation. In summary DUSt3R makes many geometric 3D vision tasks easy. Code and models at https://github.com/naver/dust3r