Minsu Cho

Doyup Lee, Chiheon Kim, Saehoon Kim et al.

For autoregressive (AR) modeling of high-resolution images, vector quantization (VQ) represents an image as a sequence of discrete codes. A short sequence length is important for an AR model to reduce its computational costs to consider long-range interactions of codes. However, we postulate that previous VQ cannot shorten the code sequence and generate high-fidelity images together in terms of the rate-distortion trade-off. In this study, we propose the two-stage framework, which consists of Residual-Quantized VAE (RQ-VAE) and RQ-Transformer, to effectively generate high-resolution images. Given a fixed codebook size, RQ-VAE can precisely approximate a feature map of an image and represent the image as a stacked map of discrete codes. Then, RQ-Transformer learns to predict the quantized feature vector at the next position by predicting the next stack of codes. Thanks to the precise approximation of RQ-VAE, we can represent a 256$\times$256 image as 8$\times$8 resolution of the feature map, and RQ-Transformer can efficiently reduce the computational costs. Consequently, our framework outperforms the existing AR models on various benchmarks of unconditional and conditional image generation. Our approach also has a significantly faster sampling speed than previous AR models to generate high-quality images.

Dayoung Gong, Joonseok Lee, Manjin Kim et al.

The task of predicting future actions from a video is crucial for a real-world agent interacting with others. When anticipating actions in the distant future, we humans typically consider long-term relations over the whole sequence of actions, i.e., not only observed actions in the past but also potential actions in the future. In a similar spirit, we propose an end-to-end attention model for action anticipation, dubbed Future Transformer (FUTR), that leverages global attention over all input frames and output tokens to predict a minutes-long sequence of future actions. Unlike the previous autoregressive models, the proposed method learns to predict the whole sequence of future actions in parallel decoding, enabling more accurate and fast inference for long-term anticipation. We evaluate our method on two standard benchmarks for long-term action anticipation, Breakfast and 50 Salads, achieving state-of-the-art results.

Dahyun Kang, Minsu Cho

We introduce the integrative task of few-shot classification and segmentation (FS-CS) that aims to both classify and segment target objects in a query image when the target classes are given with a few examples. This task combines two conventional few-shot learning problems, few-shot classification and segmentation. FS-CS generalizes them to more realistic episodes with arbitrary image pairs, where each target class may or may not be present in the query. To address the task, we propose the integrative few-shot learning (iFSL) framework for FS-CS, which trains a learner to construct class-wise foreground maps for multi-label classification and pixel-wise segmentation. We also develop an effective iFSL model, attentive squeeze network (ASNet), that leverages deep semantic correlation and global self-attention to produce reliable foreground maps. In experiments, the proposed method shows promising performance on the FS-CS task and also achieves the state of the art on standard few-shot segmentation benchmarks.

Minji Kim, Seungkwan Lee, Jungseul Ok et al.

Despite the extensive adoption of machine learning on the task of visual object tracking, recent learning-based approaches have largely overlooked the fact that visual tracking is a sequence-level task in its nature; they rely heavily on frame-level training, which inevitably induces inconsistency between training and testing in terms of both data distributions and task objectives. This work introduces a sequence-level training strategy for visual tracking based on reinforcement learning and discusses how a sequence-level design of data sampling, learning objectives, and data augmentation can improve the accuracy and robustness of tracking algorithms. Our experiments on standard benchmarks including LaSOT, TrackingNet, and GOT-10k demonstrate that four representative tracking models, SiamRPN++, SiamAttn, TransT, and TrDiMP, consistently improve by incorporating the proposed methods in training without modifying architectures.

Sanghyun Kim, Deunsol Jung, Minsu Cho

Recent state-of-the-art methods for HOI detection typically build on transformer architectures with two decoder branches, one for human-object pair detection and the other for interaction classification. Such disentangled transformers, however, may suffer from insufficient context exchange between the branches and lead to a lack of context information for relational reasoning, which is critical in discovering HOI instances. In this work, we propose the multiplex relation network (MUREN) that performs rich context exchange between three decoder branches using unary, pairwise, and ternary relations of human, object, and interaction tokens. The proposed method learns comprehensive relational contexts for discovering HOI instances, achieving state-of-the-art performance on two standard benchmarks for HOI detection, HICO-DET and V-COCO.

Dongkeun Kim, Jinsung Lee, Minsu Cho et al.

Group activity recognition is the task of understanding the activity conducted by a group of people as a whole in a multi-person video. Existing models for this task are often impractical in that they demand ground-truth bounding box labels of actors even in testing or rely on off-the-shelf object detectors. Motivated by this, we propose a novel model for group activity recognition that depends neither on bounding box labels nor on object detector. Our model based on Transformer localizes and encodes partial contexts of a group activity by leveraging the attention mechanism, and represents a video clip as a set of partial context embeddings. The embedding vectors are then aggregated to form a single group representation that reflects the entire context of an activity while capturing temporal evolution of each partial context. Our method achieves outstanding performance on two benchmarks, Volleyball and NBA datasets, surpassing not only the state of the art trained with the same level of supervision, but also some of existing models relying on stronger supervision.

Seungwook Kim, Juhong Min, Minsu Cho

Establishing correspondences between images remains a challenging task, especially under large appearance changes due to different viewpoints or intra-class variations. In this work, we introduce a strong semantic image matching learner, dubbed TransforMatcher, which builds on the success of transformer networks in vision domains. Unlike existing convolution- or attention-based schemes for correspondence, TransforMatcher performs global match-to-match attention for precise match localization and dynamic refinement. To handle a large number of matches in a dense correlation map, we develop a light-weight attention architecture to consider the global match-to-match interactions. We also propose to utilize a multi-channel correlation map for refinement, treating the multi-level scores as features instead of a single score to fully exploit the richer layer-wise semantics. In experiments, TransforMatcher sets a new state of the art on SPair-71k while performing on par with existing SOTA methods on the PF-PASCAL dataset.

Jongmin Lee, Byungjin Kim, Minsu Cho

Detecting robust keypoints from an image is an integral part of many computer vision problems, and the characteristic orientation and scale of keypoints play an important role for keypoint description and matching. Existing learning-based methods for keypoint detection rely on standard translation-equivariant CNNs but often fail to detect reliable keypoints against geometric variations. To learn to detect robust oriented keypoints, we introduce a self-supervised learning framework using rotation-equivariant CNNs. We propose a dense orientation alignment loss by an image pair generated by synthetic transformations for training a histogram-based orientation map. Our method outperforms the previous methods on an image matching benchmark and a camera pose estimation benchmark.

Deunsol Jung, Sanghyun Kim, Won Hwa Kim et al.

Scene graph generation aims to construct a semantic graph structure from an image such that its nodes and edges respectively represent objects and their relationships. One of the major challenges for the task lies in the presence of distracting objects and relationships in images; contextual reasoning is strongly distracted by irrelevant objects or backgrounds and, more importantly, a vast number of irrelevant candidate relations. To tackle the issue, we propose the Selective Quad Attention Network (SQUAT) that learns to select relevant object pairs and disambiguate them via diverse contextual interactions. SQUAT consists of two main components: edge selection and quad attention. The edge selection module selects relevant object pairs, i.e., edges in the scene graph, which helps contextual reasoning, and the quad attention module then updates the edge features using both edge-to-node and edge-to-edge cross-attentions to capture contextual information between objects and object pairs. Experiments demonstrate the strong performance and robustness of SQUAT, achieving the state of the art on the Visual Genome and Open Images v6 benchmarks.

Jongmin Lee, Byungjin Kim, Seungwook Kim et al.

Extracting discriminative local features that are invariant to imaging variations is an integral part of establishing correspondences between images. In this work, we introduce a self-supervised learning framework to extract discriminative rotation-invariant descriptors using group-equivariant CNNs. Thanks to employing group-equivariant CNNs, our method effectively learns to obtain rotation-equivariant features and their orientations explicitly, without having to perform sophisticated data augmentations. The resultant features and their orientations are further processed by group aligning, a novel invariant mapping technique that shifts the group-equivariant features by their orientations along the group dimension. Our group aligning technique achieves rotation-invariance without any collapse of the group dimension and thus eschews loss of discriminability. The proposed method is trained end-to-end in a self-supervised manner, where we use an orientation alignment loss for the orientation estimation and a contrastive descriptor loss for robust local descriptors to geometric/photometric variations. Our method demonstrates state-of-the-art matching accuracy among existing rotation-invariant descriptors under varying rotation and also shows competitive results when transferred to the task of keypoint matching and camera pose estimation.

Juhong Min, Yucheng Zhao, Chong Luo et al.

Human vision possesses a special type of visual processing systems called peripheral vision. Partitioning the entire visual field into multiple contour regions based on the distance to the center of our gaze, the peripheral vision provides us the ability to perceive various visual features at different regions. In this work, we take a biologically inspired approach and explore to model peripheral vision in deep neural networks for visual recognition. We propose to incorporate peripheral position encoding to the multi-head self-attention layers to let the network learn to partition the visual field into diverse peripheral regions given training data. We evaluate the proposed network, dubbed PerViT, on ImageNet-1K and systematically investigate the inner workings of the model for machine perception, showing that the network learns to perceive visual data similarly to the way that human vision does. The performance improvements in image classification over the baselines across different model sizes demonstrate the efficacy of the proposed method.

Naren Dhyani, Jianqiao Mo, Minsu Cho et al.

The Vision Transformer (ViT) architecture has emerged as the backbone of choice for state-of-the-art deep models for computer vision applications. However, ViTs are ill-suited for private inference using secure multi-party computation (MPC) protocols, due to the large number of non-polynomial operations (self-attention, feed-forward rectifiers, layer normalization). We propose PriViT, a gradient based algorithm to selectively "Taylorize" nonlinearities in ViTs while maintaining their prediction accuracy. Our algorithm is conceptually simple, easy to implement, and achieves improved performance over existing approaches for designing MPC-friendly transformer architectures in terms of achieving the Pareto frontier in latency-accuracy. We confirm these improvements via experiments on several standard image classification tasks. Public code is available at https://github.com/NYU-DICE-Lab/privit.

Dahyun Kang, Piotr Koniusz, Minsu Cho et al.

We address the task of weakly-supervised few-shot image classification and segmentation, by leveraging a Vision Transformer (ViT) pretrained with self-supervision. Our proposed method takes token representations from the self-supervised ViT and leverages their correlations, via self-attention, to produce classification and segmentation predictions through separate task heads. Our model is able to effectively learn to perform classification and segmentation in the absence of pixel-level labels during training, using only image-level labels. To do this it uses attention maps, created from tokens generated by the self-supervised ViT backbone, as pixel-level pseudo-labels. We also explore a practical setup with ``mixed" supervision, where a small number of training images contains ground-truth pixel-level labels and the remaining images have only image-level labels. For this mixed setup, we propose to improve the pseudo-labels using a pseudo-label enhancer that was trained using the available ground-truth pixel-level labels. Experiments on Pascal-5i and COCO-20i demonstrate significant performance gains in a variety of supervision settings, and in particular when little-to-no pixel-level labels are available.

Doyup Lee, Chiheon Kim, Saehoon Kim et al.

Although autoregressive models have achieved promising results on image generation, their unidirectional generation process prevents the resultant images from fully reflecting global contexts. To address the issue, we propose an effective image generation framework of Draft-and-Revise with Contextual RQ-transformer to consider global contexts during the generation process. As a generalized VQ-VAE, RQ-VAE first represents a high-resolution image as a sequence of discrete code stacks. After code stacks in the sequence are randomly masked, Contextual RQ-Transformer is trained to infill the masked code stacks based on the unmasked contexts of the image. Then, Contextual RQ-Transformer uses our two-phase decoding, Draft-and-Revise, and generates an image, while exploiting the global contexts of the image during the generation process. Specifically. in the draft phase, our model first focuses on generating diverse images despite rather low quality. Then, in the revise phase, the model iteratively improves the quality of images, while preserving the global contexts of generated images. In experiments, our method achieves state-of-the-art results on conditional image generation. We also validate that the Draft-and-Revise decoding can achieve high performance by effectively controlling the quality-diversity trade-off in image generation.

Chiheon Kim, Doyup Lee, Saehoon Kim et al.

Despite recent advances in implicit neural representations (INRs), it remains challenging for a coordinate-based multi-layer perceptron (MLP) of INRs to learn a common representation across data instances and generalize it for unseen instances. In this work, we introduce a simple yet effective framework for generalizable INRs that enables a coordinate-based MLP to represent complex data instances by modulating only a small set of weights in an early MLP layer as an instance pattern composer; the remaining MLP weights learn pattern composition rules for common representations across instances. Our generalizable INR framework is fully compatible with existing meta-learning and hypernetworks in learning to predict the modulated weight for unseen instances. Extensive experiments demonstrate that our method achieves high performance on a wide range of domains such as an audio, image, and 3D object, while the ablation study validates our weight modulation.

Minh Pham, Minsu Cho, Ameya Joshi et al.

Knowledge distillation is the procedure of transferring "knowledge" from a large model (the teacher) to a more compact one (the student), often being used in the context of model compression. When both models have the same architecture, this procedure is called self-distillation. Several works have anecdotally shown that a self-distilled student can outperform the teacher on held-out data. In this work, we systematically study self-distillation in a number of settings. We first show that even with a highly accurate teacher, self-distillation allows a student to surpass the teacher in all cases. Secondly, we revisit existing theoretical explanations of (self) distillation and identify contradicting examples, revealing possible drawbacks of these explanations. Finally, we provide an alternative explanation for the dynamics of self-distillation through the lens of loss landscape geometry. We conduct extensive experiments to show that self-distillation leads to flatter minima, thereby resulting in better generalization.

Hyolim Kang, Hanjung Kim, Joungbin An et al.

Temporal Action Localization (TAL) methods typically operate on top of feature sequences from a frozen snippet encoder that is pretrained with the Trimmed Action Classification (TAC) tasks, resulting in a task discrepancy problem. While existing TAL methods mitigate this issue either by retraining the encoder with a pretext task or by end-to-end fine-tuning, they commonly require an overload of high memory and computation. In this work, we introduce Soft-Landing (SoLa) strategy, an efficient yet effective framework to bridge the transferability gap between the pretrained encoder and the downstream tasks by incorporating a light-weight neural network, i.e., a SoLa module, on top of the frozen encoder. We also propose an unsupervised training scheme for the SoLa module; it learns with inter-frame Similarity Matching that uses the frame interval as its supervisory signal, eliminating the need for temporal annotations. Experimental evaluation on various benchmarks for downstream TAL tasks shows that our method effectively alleviates the task discrepancy problem with remarkable computational efficiency.

Jinhwi Lee, Jungtaek Kim, Hyunsoo Chung et al.

Assembling parts into an object is a combinatorial problem that arises in a variety of contexts in the real world and involves numerous applications in science and engineering. Previous related work tackles limited cases with identical unit parts or jigsaw-style parts of textured shapes, which greatly mitigate combinatorial challenges of the problem. In this work, we introduce the more challenging problem of shape assembly, which involves textureless fragments of arbitrary shapes with indistinctive junctions, and then propose a learning-based approach to solving it. We demonstrate the effectiveness on shape assembly tasks with various scenarios, including the ones with abnormal fragments (e.g., missing and distorted), the different number of fragments, and different rotation discretization.

Seungwook Kim, Chunghyun Park, Yoonwoo Jeong et al.

Learning to predict reliable characteristic orientations of 3D point clouds is an important yet challenging problem, as different point clouds of the same class may have largely varying appearances. In this work, we introduce a novel method to decouple the shape geometry and semantics of the input point cloud to achieve both stability and consistency. The proposed method integrates shape-geometry-based SO(3)-equivariant learning and shape-semantics-based SO(3)-invariant residual learning, where a final characteristic orientation is obtained by calibrating an SO(3)-equivariant orientation hypothesis using an SO(3)-invariant residual rotation. In experiments, the proposed method not only demonstrates superior stability and consistency but also exhibits state-of-the-art performances when applied to point cloud part segmentation, given randomly rotated inputs.

Seokjun Ahn, Jungtaek Kim, Minsu Cho et al.

We tackle the problem of sequential brick assembly with LEGO bricks to create combinatorial 3D structures. This problem is challenging since this brick assembly task encompasses the characteristics of combinatorial optimization problems. In particular, the number of assemblable structures increases exponentially as the number of bricks used increases. To solve this problem, we propose a new method to predict the scores of the next brick position by employing a U-shaped sparse 3D convolutional neural network. Along with the 3D convolutional network, a one-initialized brick-sized convolution filter is used to efficiently validate assembly constraints between bricks without training itself. By the nature of this one-initialized convolution filter, we can readily consider several different brick types by benefiting from modern implementation of convolution operations. To generate a novel structure, we devise a sampling strategy to determine the next brick position considering the satisfaction of assembly constraints. Moreover, our method is designed for either budget-free or budget-aware scenario where a budget may confine the number of bricks and their types. We demonstrate that our method successfully generates a variety of brick structures and outperforms existing methods with Bayesian optimization, deep graph generative model, and reinforcement learning.

Jinsung Lee, Taeoh Kim, Inwoong Lee et al.

Video action detection (VAD) aims to detect actors and classify their actions in a video. We figure that VAD suffers more from classification rather than localization of actors. Hence, we analyze how prevailing methods form features for classification and find that they prioritize actor regions, yet often overlooking the essential contextual information necessary for accurate classification. Accordingly, we propose to reduce the bias toward actor and encourage paying attention to the context that is relevant to each action class. By assigning a class-dedicated query to each action class, our model can dynamically determine where to focus for effective classification. The proposed model demonstrates superior performance on three challenging benchmarks with significantly fewer parameters and less computation.

Doyup Lee, Chiheon Kim, Minsu Cho et al.

Generalizable implicit neural representation (INR) enables a single continuous function, i.e., a coordinate-based neural network, to represent multiple data instances by modulating its weights or intermediate features using latent codes. However, the expressive power of the state-of-the-art modulation is limited due to its inability to localize and capture fine-grained details of data entities such as specific pixels and rays. To address this issue, we propose a novel framework for generalizable INR that combines a transformer encoder with a locality-aware INR decoder. The transformer encoder predicts a set of latent tokens from a data instance to encode local information into each latent token. The locality-aware INR decoder extracts a modulation vector by selectively aggregating the latent tokens via cross-attention for a coordinate input and then predicts the output by progressively decoding with coarse-to-fine modulation through multiple frequency bandwidths. The selective token aggregation and the multi-band feature modulation enable us to learn locality-aware representation in spatial and spectral aspects, respectively. Our framework significantly outperforms previous generalizable INRs and validates the usefulness of the locality-aware latents for downstream tasks such as image generation.

Jongmin Lee, Yoonwoo Jeong, Minsu Cho

We study the problem of learning to assign a characteristic pose, i.e., scale and orientation, for an image region of interest. Despite its apparent simplicity, the problem is non-trivial; it is hard to obtain a large-scale set of image regions with explicit pose annotations that a model directly learns from. To tackle the issue, we propose a self-supervised learning framework with a histogram alignment technique. It generates pairs of image patches by random rescaling/rotating and then train an estimator to predict their scale/orientation values so that their relative difference is consistent with the rescaling/rotating used. The estimator learns to predict a non-parametric histogram distribution of scale/orientation without any supervision. Experiments show that it significantly outperforms previous methods in scale/orientation estimation and also improves image matching and 6 DoF camera pose estimation by incorporating our patch poses into a matching process.

Nahyuk Lee, Juhong Min, Junha Lee et al.

Learning to assemble geometric shapes into a larger target structure is a pivotal task in various practical applications. In this work, we tackle this problem by establishing local correspondences between point clouds of part shapes in both coarse- and fine-levels. To this end, we introduce Proxy Match Transform (PMT), an approximate high-order feature transform layer that enables reliable matching between mating surfaces of parts while incurring low costs in memory and computation. Building upon PMT, we introduce a new framework, dubbed Proxy Match TransformeR (PMTR), for the geometric assembly task. We evaluate the proposed PMTR on the large-scale 3D geometric shape assembly benchmark dataset of Breaking Bad and demonstrate its superior performance and efficiency compared to state-of-the-art methods. Project page: https://nahyuklee.github.io/pmtr.

Yoonwoo Jeong, Seungjoo Shin, Junha Lee et al.

The recent progress in implicit 3D representation, i.e., Neural Radiance Fields (NeRFs), has made accurate and photorealistic 3D reconstruction possible in a differentiable manner. This new representation can effectively convey the information of hundreds of high-resolution images in one compact format and allows photorealistic synthesis of novel views. In this work, using the variant of NeRF called Plenoxels, we create the first large-scale implicit representation datasets for perception tasks, called the PeRFception, which consists of two parts that incorporate both object-centric and scene-centric scans for classification and segmentation. It shows a significant memory compression rate (96.4\%) from the original dataset, while containing both 2D and 3D information in a unified form. We construct the classification and segmentation models that directly take as input this implicit format and also propose a novel augmentation technique to avoid overfitting on backgrounds of images. The code and data are publicly available in https://postech-cvlab.github.io/PeRFception .

Jungtaek Kim, Jeongbeen Yoon, Minsu Cho

Sorting is a fundamental operation of all computer systems, having been a long-standing significant research topic. Beyond the problem formulation of traditional sorting algorithms, we consider sorting problems for more abstract yet expressive inputs, e.g., multi-digit images and image fragments, through a neural sorting network. To learn a mapping from a high-dimensional input to an ordinal variable, the differentiability of sorting networks needs to be guaranteed. In this paper we define a softening error by a differentiable swap function, and develop an error-free swap function that holds a non-decreasing condition and differentiability. Furthermore, a permutation-equivariant Transformer network with multi-head attention is adopted to capture dependency between given inputs and also leverage its model capacity with self-attention. Experiments on diverse sorting benchmarks show that our methods perform better than or comparable to baseline methods.

Junha Lee, Eunha Park, Chunghyun Park et al.

Affordance grounding-localizing object regions based on natural language descriptions of interactions-is a critical challenge for enabling intelligent agents to understand and interact with their environments. However, this task remains challenging due to the need for fine-grained part-level localization, the ambiguity arising from multiple valid interaction regions, and the scarcity of large-scale datasets. In this work, we introduce Affogato, a large-scale benchmark comprising 150K instances, annotated with open-vocabulary text descriptions and corresponding 3D affordance heatmaps across a diverse set of objects and interactions. Building on this benchmark, we develop simple yet effective vision-language models that leverage pretrained part-aware vision backbones and a text-conditional heatmap decoder. Our models trained with the Affogato dataset achieve promising performance on the existing 2D and 3D benchmarks, and notably, exhibit effectiveness in open-vocabulary cross-domain generalization. The Affogato dataset is shared in public: https://huggingface.co/datasets/project-affogato/affogato

Minsu Cho, Ameya Joshi, Siddharth Garg et al.

Private inference (PI) enables inference directly on cryptographically secure data.While promising to address many privacy issues, it has seen limited use due to extreme runtimes. Unlike plaintext inference, where latency is dominated by FLOPs, in PI non-linear functions (namely ReLU) are the bottleneck. Thus, practical PI demands novel ReLU-aware optimizations. To reduce PI latency we propose a gradient-based algorithm that selectively linearizes ReLUs while maintaining prediction accuracy. We evaluate our algorithm on several standard PI benchmarks. The results demonstrate up to $4.25\%$ more accuracy (iso-ReLU count at 50K) or $2.2\times$ less latency (iso-accuracy at 70\%) than the current state of the art and advance the Pareto frontier across the latency-accuracy space. To complement empirical results, we present a "no free lunch" theorem that sheds light on how and when network linearization is possible while maintaining prediction accuracy. Public code is available at \url{https://github.com/NYU-DICE-Lab/selective_network_linearization}.

Gyeongsik Moon, Heeseung Kwon, Kyoung Mu Lee et al.

Most current action recognition methods heavily rely on appearance information by taking an RGB sequence of entire image regions as input. While being effective in exploiting contextual information around humans, e.g., human appearance and scene category, they are easily fooled by out-of-context action videos where the contexts do not exactly match with target actions. In contrast, pose-based methods, which take a sequence of human skeletons only as input, suffer from inaccurate pose estimation or ambiguity of human pose per se. Integrating these two approaches has turned out to be non-trivial; training a model with both appearance and pose ends up with a strong bias towards appearance and does not generalize well to unseen videos. To address this problem, we propose to learn pose-driven feature integration that dynamically combines appearance and pose streams by observing pose features on the fly. The main idea is to let the pose stream decide how much and which appearance information is used in integration based on whether the given pose information is reliable or not. We show that the proposed IntegralAction achieves highly robust performance across in-context and out-of-context action video datasets. The codes are available in https://github.com/mks0601/IntegralAction_RELEASE.

Juhong Min, Shyamal Buch, Arsha Nagrani et al.

This paper addresses the task of video question answering (videoQA) via a decomposed multi-stage, modular reasoning framework. Previous modular methods have shown promise with a single planning stage ungrounded in visual content. However, through a simple and effective baseline, we find that such systems can lead to brittle behavior in practice for challenging videoQA settings. Thus, unlike traditional single-stage planning methods, we propose a multi-stage system consisting of an event parser, a grounding stage, and a final reasoning stage in conjunction with an external memory. All stages are training-free, and performed using few-shot prompting of large models, creating interpretable intermediate outputs at each stage. By decomposing the underlying planning and task complexity, our method, MoReVQA, improves over prior work on standard videoQA benchmarks (NExT-QA, iVQA, EgoSchema, ActivityNet-QA) with state-of-the-art results, and extensions to related tasks (grounded videoQA, paragraph captioning).

Manjin Kim, Paul Hongsuck Seo, Cordelia Schmid et al.

We introduce a new attention mechanism, dubbed structural self-attention (StructSA), that leverages rich correlation patterns naturally emerging in key-query interactions of attention. StructSA generates attention maps by recognizing space-time structures of key-query correlations via convolution and uses them to dynamically aggregate local contexts of value features. This effectively leverages rich structural patterns in images and videos such as scene layouts, object motion, and inter-object relations. Using StructSA as a main building block, we develop the structural vision transformer (StructViT) and evaluate its effectiveness on both image and video classification tasks, achieving state-of-the-art results on ImageNet-1K, Kinetics-400, Something-Something V1 & V2, Diving-48, and FineGym.

Dayoung Gong, Joonseok Lee, Deunsol Jung et al.

Sequence prediction on temporal data requires the ability to understand compositional structures of multi-level semantics beyond individual and contextual properties. The task of temporal action segmentation, which aims at translating an untrimmed activity video into a sequence of action segments, remains challenging for this reason. This paper addresses the problem by introducing an effective activity grammar to guide neural predictions for temporal action segmentation. We propose a novel grammar induction algorithm that extracts a powerful context-free grammar from action sequence data. We also develop an efficient generalized parser that transforms frame-level probability distributions into a reliable sequence of actions according to the induced grammar with recursive rules. Our approach can be combined with any neural network for temporal action segmentation to enhance the sequence prediction and discover its compositional structure. Experimental results demonstrate that our method significantly improves temporal action segmentation in terms of both performance and interpretability on two standard benchmarks, Breakfast and 50 Salads.

Yoonwoo Jeong, Jinwoo Lee, Chiheon Kim et al.

Transfer learning of large-scale Text-to-Image (T2I) models has recently shown impressive potential for Novel View Synthesis (NVS) of diverse objects from a single image. While previous methods typically train large models on multi-view datasets for NVS, fine-tuning the whole parameters of T2I models not only demands a high cost but also reduces the generalization capacity of T2I models in generating diverse images in a new domain. In this study, we propose an effective method, dubbed NVS-Adapter, which is a plug-and-play module for a T2I model, to synthesize novel multi-views of visual objects while fully exploiting the generalization capacity of T2I models. NVS-Adapter consists of two main components; view-consistency cross-attention learns the visual correspondences to align the local details of view features, and global semantic conditioning aligns the semantic structure of generated views with the reference view. Experimental results demonstrate that the NVS-Adapter can effectively synthesize geometrically consistent multi-views and also achieve high performance on benchmarks without full fine-tuning of T2I models. The code and data are publicly available in ~\href{https://postech-cvlab.github.io/nvsadapter/}{https://postech-cvlab.github.io/nvsadapter/}.

Seungwook Kim, Yichun Shi, Kejie Li et al.

Using image as prompts for 3D generation demonstrate particularly strong performances compared to using text prompts alone, for images provide a more intuitive guidance for the 3D generation process. In this work, we delve into the potential of using multiple image prompts, instead of a single image prompt, for 3D generation. Specifically, we build on ImageDream, a novel image-prompt multi-view diffusion model, to support multi-view images as the input prompt. Our method, dubbed MultiImageDream, reveals that transitioning from a single-image prompt to multiple-image prompts enhances the performance of multi-view and 3D object generation according to various quantitative evaluation metrics and qualitative assessments. This advancement is achieved without the necessity of fine-tuning the pre-trained ImageDream multi-view diffusion model.

Seungwook Kim, Kejie Li, Xueqing Deng et al.

Leveraging multi-view diffusion models as priors for 3D optimization have alleviated the problem of 3D consistency, e.g., the Janus face problem or the content drift problem, in zero-shot text-to-3D models. However, the 3D geometric fidelity of the output remains an unresolved issue; albeit the rendered 2D views are realistic, the underlying geometry may contain errors such as unreasonable concavities. In this work, we propose CorrespondentDream, an effective method to leverage annotation-free, cross-view correspondences yielded from the diffusion U-Net to provide additional 3D prior to the NeRF optimization process. We find that these correspondences are strongly consistent with human perception, and by adopting it in our loss design, we are able to produce NeRF models with geometries that are more coherent with common sense, e.g., more smoothed object surface, yielding higher 3D fidelity. We demonstrate the efficacy of our approach through various comparative qualitative results and a solid user study.

Junha Lee, Chunghyun Park, Jaesung Choe et al.

We tackle open-vocabulary 3D scene understanding by introducing a novel data generation pipeline and training framework. Our method addresses three critical requirements for effective training: precise 3D region segmentation, comprehensive textual descriptions, and sufficient dataset scale. By leveraging state-of-the-art open-vocabulary image segmentation models and region-aware Vision-Language Models, we develop an automatic pipeline that generates high-quality 3D mask-text pairs. Applying this pipeline to multiple 3D scene datasets, we create Mosaic3D-5.6M, a dataset of over 30K annotated scenes with 5.6M mask-text pairs, significantly larger than existing datasets. Building upon this data, we propose Mosaic3D, a foundation model combining a 3D encoder trained with contrastive learning and a lightweight mask decoder for open-vocabulary 3D semantic and instance segmentation. Our approach achieves state-of-the-art results on open-vocabulary 3D semantic and instance segmentation tasks including ScanNet200, Matterport3D, and ScanNet++, with ablation studies validating the effectiveness of our large-scale training data.

Chunghyun Park, Seungwook Kim, Jaesik Park et al.

Establishing accurate 3D correspondences between shapes stands as a pivotal challenge with profound implications for computer vision and robotics. However, existing self-supervised methods for this problem assume perfect input shape alignment, restricting their real-world applicability. In this work, we introduce a novel self-supervised Rotation-Invariant 3D correspondence learner with Local Shape Transform, dubbed RIST, that learns to establish dense correspondences between shapes even under challenging intra-class variations and arbitrary orientations. Specifically, RIST learns to dynamically formulate an SO(3)-invariant local shape transform for each point, which maps the SO(3)-equivariant global shape descriptor of the input shape to a local shape descriptor. These local shape descriptors are provided as inputs to our decoder to facilitate point cloud self- and cross-reconstruction. Our proposed self-supervised training pipeline encourages semantically corresponding points from different shapes to be mapped to similar local shape descriptors, enabling RIST to establish dense point-wise correspondences. RIST demonstrates state-of-the-art performances on 3D part label transfer and semantic keypoint transfer given arbitrarily rotated point cloud pairs, outperforming existing methods by significant margins.

Ahyun Seo, Minsu Cho

Symmetry plays a vital role in understanding structural patterns, aiding object recognition and scene interpretation. This paper focuses on rotation symmetry, where objects remain unchanged when rotated around a central axis, requiring detection of rotation centers and supporting vertices. Traditional methods relied on hand-crafted feature matching, while recent segmentation models based on convolutional neural networks detect rotation centers but struggle with 3D geometric consistency due to viewpoint distortions. To overcome this, we propose a model that directly predicts rotation centers and vertices in 3D space and projects the results back to 2D while preserving structural integrity. By incorporating a vertex reconstruction stage enforcing 3D geometric priors -- such as equal side lengths and interior angles -- our model enhances robustness and accuracy. Experiments on the DENDI dataset show superior performance in rotation axis detection and validate the impact of 3D priors through ablation studies.

Seungwook Kim, Minsu Cho

Establishing semantic correspondences between images is a fundamental yet challenging task in computer vision. Traditional feature-metric methods enhance visual features but may miss complex inter-correlation relationships, while recent correlation-metric approaches are hindered by high computational costs due to processing 4D correlation maps. We introduce MambaMatcher, a novel method that overcomes these limitations by efficiently modeling high-dimensional correlations using selective state-space models (SSMs). By implementing a similarity-aware selective scan mechanism adapted from Mamba's linear-complexity algorithm, MambaMatcher refines the 4D correlation map effectively without compromising feature map resolution or receptive field. Experiments on standard semantic correspondence benchmarks demonstrate that MambaMatcher achieves state-of-the-art performance.

Sanghyun Kim, Deunsol Jung, Minsu Cho

Recent methods for zero-shot Human-Object Interaction (HOI) detection typically leverage the generalization ability of large Vision-Language Model (VLM), i.e., CLIP, on unseen categories, showing impressive results on various zero-shot settings. However, existing methods struggle to adapt CLIP representations for human-object pairs, as CLIP tends to overlook fine-grained information necessary for distinguishing interactions. To address this issue, we devise, LAIN, a novel zero-shot HOI detection framework enhancing the locality and interaction awareness of CLIP representations. The locality awareness, which involves capturing fine-grained details and the spatial structure of individual objects, is achieved by aggregating the information and spatial priors of adjacent neighborhood patches. The interaction awareness, which involves identifying whether and how a human is interacting with an object, is achieved by capturing the interaction pattern between the human and the object. By infusing locality and interaction awareness into CLIP representation, LAIN captures detailed information about the human-object pairs. Our extensive experiments on existing benchmarks show that LAIN outperforms previous methods on various zero-shot settings, demonstrating the importance of locality and interaction awareness for effective zero-shot HOI detection.

Jongseong Bae, Susang Kim, Minsu Cho et al.

Active research is currently underway to enhance the efficiency of vision transformers (ViTs). Most studies have focused solely on effective token mixers, overlooking the potential relationship with normalization. To boost diverse feature learning, we propose two components: a normalization module called multi-view normalization (MVN) and a token mixer called multi-view token mixer (MVTM). The MVN integrates three differently normalized features via batch, layer, and instance normalization using a learnable weighted sum. Each normalization method outputs a different distribution, generating distinct features. Thus, the MVN is expected to offer diverse pattern information to the token mixer, resulting in beneficial synergy. The MVTM is a convolution-based multiscale token mixer with local, intermediate, and global filters, and it incorporates stage specificity by configuring various receptive fields for the token mixer at each stage, efficiently capturing ranges of visual patterns. We propose a novel ViT model, multi-vision transformer (MVFormer), adopting the MVN and MVTM in the MetaFormer block, the generalized ViT scheme. Our MVFormer outperforms state-of-the-art convolution-based ViTs on image classification, object detection, and instance and semantic segmentation with the same or lower parameters and MACs. Particularly, MVFormer variants, MVFormer-T, S, and B achieve 83.4%, 84.3%, and 84.6% top-1 accuracy, respectively, on ImageNet-1K benchmark.

Ahyun Seo, Byungjin Kim, Suha Kwak et al.

The inherent challenge of detecting symmetries stems from arbitrary orientations of symmetry patterns; a reflection symmetry mirrors itself against an axis with a specific orientation while a rotation symmetry matches its rotated copy with a specific orientation. Discovering such symmetry patterns from an image thus benefits from an equivariant feature representation, which varies consistently with reflection and rotation of the image. In this work, we introduce a group-equivariant convolutional network for symmetry detection, dubbed EquiSym, which leverages equivariant feature maps with respect to a dihedral group of reflection and rotation. The proposed network is built end-to-end with dihedrally-equivariant layers and trained to output a spatial map for reflection axes or rotation centers. We also present a new dataset, DENse and DIverse symmetry (DENDI), which mitigates limitations of existing benchmarks for reflection and rotation symmetry detection. Experiments show that our method achieves the state of the arts in symmetry detection on LDRS and DENDI datasets.

Doyup Lee, Sungwoong Kim, Ildoo Kim et al.

Consistency regularization on label predictions becomes a fundamental technique in semi-supervised learning, but it still requires a large number of training iterations for high performance. In this study, we analyze that the consistency regularization restricts the propagation of labeling information due to the exclusion of samples with unconfident pseudo-labels in the model updates. Then, we propose contrastive regularization to improve both efficiency and accuracy of the consistency regularization by well-clustered features of unlabeled data. In specific, after strongly augmented samples are assigned to clusters by their pseudo-labels, our contrastive regularization updates the model so that the features with confident pseudo-labels aggregate the features in the same cluster, while pushing away features in different clusters. As a result, the information of confident pseudo-labels can be effectively propagated into more unlabeled samples during training by the well-clustered features. On benchmarks of semi-supervised learning tasks, our contrastive regularization improves the previous consistency-based methods and achieves state-of-the-art results, especially with fewer training iterations. Our method also shows robust performance on open-set semi-supervised learning where unlabeled data includes out-of-distribution samples.

Jeongbeen Yoon, Dahyun Kang, Minsu Cho

Semi-supervised domain adaptation (SSDA) is to adapt a learner to a new domain with only a small set of labeled samples when a large labeled dataset is given on a source domain. In this paper, we propose a pair-based SSDA method that adapts a model to the target domain using self-distillation with sample pairs. Each sample pair is composed of a teacher sample from a labeled dataset (i.e., source or labeled target) and its student sample from an unlabeled dataset (i.e., unlabeled target). Our method generates an assistant feature by transferring an intermediate style between the teacher and the student, and then train the model by minimizing the output discrepancy between the student and the assistant. During training, the assistants gradually bridge the discrepancy between the two domains, thus allowing the student to easily learn from the teacher. Experimental evaluation on standard benchmarks shows that our method effectively minimizes both the inter-domain and intra-domain discrepancies, thus achieving significant improvements over recent methods.

Juhong Min, Seungwook Kim, Minsu Cho

Despite advances in feature representation, leveraging geometric relations is crucial for establishing reliable visual correspondences under large variations of images. In this work we introduce a Hough transform perspective on convolutional matching and propose an effective geometric matching algorithm, dubbed Convolutional Hough Matching (CHM). The method distributes similarities of candidate matches over a geometric transformation space and evaluates them in a convolutional manner. We cast it into a trainable neural layer with a semi-isotropic high-dimensional kernel, which learns non-rigid matching with a small number of interpretable parameters. To further improve the efficiency of high-dimensional voting, we also propose to use an efficient kernel decomposition with center-pivot neighbors, which significantly sparsifies the proposed semi-isotropic kernels without performance degradation. To validate the proposed techniques, we develop the neural network with CHM layers that perform convolutional matching in the space of translation and scaling. Our method sets a new state of the art on standard benchmarks for semantic visual correspondence, proving its strong robustness to challenging intra-class variations.

Ahyun Seo, Woohyeon Shim, Minsu Cho

The task of reflection symmetry detection remains challenging due to significant variations and ambiguities of symmetry patterns in the wild. Furthermore, since the local regions are required to match in reflection for detecting a symmetry pattern, it is hard for standard convolutional networks, which are not equivariant to rotation and reflection, to learn the task. To address the issue, we introduce a new convolutional technique, dubbed the polar matching convolution, which leverages a polar feature pooling, a self-similarity encoding, and a systematic kernel design for axes of different angles. The proposed high-dimensional kernel convolution network effectively learns to discover symmetry patterns from real-world images, overcoming the limitations of standard convolution. In addition, we present a new dataset and introduce a self-supervised learning strategy by augmenting the dataset with synthesizing images. Experiments demonstrate that our method outperforms state-of-the-art methods in terms of accuracy and robustness.

Juhong Min, Dahyun Kang, Minsu Cho

Few-shot semantic segmentation aims at learning to segment a target object from a query image using only a few annotated support images of the target class. This challenging task requires to understand diverse levels of visual cues and analyze fine-grained correspondence relations between the query and the support images. To address the problem, we propose Hypercorrelation Squeeze Networks (HSNet) that leverages multi-level feature correlation and efficient 4D convolutions. It extracts diverse features from different levels of intermediate convolutional layers and constructs a collection of 4D correlation tensors, i.e., hypercorrelations. Using efficient center-pivot 4D convolutions in a pyramidal architecture, the method gradually squeezes high-level semantic and low-level geometric cues of the hypercorrelation into precise segmentation masks in coarse-to-fine manner. The significant performance improvements on standard few-shot segmentation benchmarks of PASCAL-5i, COCO-20i, and FSS-1000 verify the efficacy of the proposed method.

Juhong Min, Minsu Cho

Despite advances in feature representation, leveraging geometric relations is crucial for establishing reliable visual correspondences under large variations of images. In this work we introduce a Hough transform perspective on convolutional matching and propose an effective geometric matching algorithm, dubbed Convolutional Hough Matching (CHM). The method distributes similarities of candidate matches over a geometric transformation space and evaluate them in a convolutional manner. We cast it into a trainable neural layer with a semi-isotropic high-dimensional kernel, which learns non-rigid matching with a small number of interpretable parameters. To validate the effect, we develop the neural network with CHM layers that perform convolutional matching in the space of translation and scaling. Our method sets a new state of the art on standard benchmarks for semantic visual correspondence, proving its strong robustness to challenging intra-class variations.

Heeseung Kwon, Manjin Kim, Suha Kwak et al.

Spatio-temporal convolution often fails to learn motion dynamics in videos and thus an effective motion representation is required for video understanding in the wild. In this paper, we propose a rich and robust motion representation based on spatio-temporal self-similarity (STSS). Given a sequence of frames, STSS represents each local region as similarities to its neighbors in space and time. By converting appearance features into relational values, it enables the learner to better recognize structural patterns in space and time. We leverage the whole volume of STSS and let our model learn to extract an effective motion representation from it. The proposed neural block, dubbed SELFY, can be easily inserted into neural architectures and trained end-to-end without additional supervision. With a sufficient volume of the neighborhood in space and time, it effectively captures long-term interaction and fast motion in the video, leading to robust action recognition. Our experimental analysis demonstrates its superiority over previous methods for motion modeling as well as its complementarity to spatio-temporal features from direct convolution. On the standard action recognition benchmarks, Something-Something-V1 & V2, Diving-48, and FineGym, the proposed method achieves the state-of-the-art results.

Jungtaek Kim, Seungjin Choi, Minsu Cho

Bayesian optimization is a popular method for solving the problem of global optimization of an expensive-to-evaluate black-box function. It relies on a probabilistic surrogate model of the objective function, upon which an acquisition function is built to determine where next to evaluate the objective function. In general, Bayesian optimization with Gaussian process regression operates on a continuous space. When input variables are categorical or discrete, an extra care is needed. A common approach is to use one-hot encoded or Boolean representation for categorical variables which might yield a combinatorial explosion problem. In this paper we present a method for Bayesian optimization in a combinatorial space, which can operate well in a large combinatorial space. The main idea is to use a random mapping which embeds the combinatorial space into a convex polytope in a continuous space, on which all essential process is performed to determine a solution to the black-box optimization in the combinatorial space. We describe our combinatorial Bayesian optimization algorithm and present its regret analysis. Numerical experiments demonstrate that our method shows satisfactory performance compared to existing methods.