ICCV 2023 Papers

Unsupervised anomaly detection and localization in images is a challenging problem, leading previous methods to attempt an easier supervised one-class classification formalization. Assuming training images to be realizations of the underlying image distribution, it follows that nominal patches from these realizations will be well associated between and represented across realizations. From this, we propose Inter-Realization Channels (InReaCh), a fully unsupervised method of detecting and localizing anomalies. InReaCh extracts high-confidence nominal patches from training data by associating them between realizations into channels, only considering channels with high spans and low spread as nominal. We then create our nominal model from the patches of these channels to test new patches against. InReaCh extracts nominal patches from the MVTec AD dataset with 99.9% precision, then archives 0.968 AUROC in localization and 0.923 AUROC in detection with corrupted training data, competitive with current state-of-the-art supervised one-class classification methods. We test our model up to 40% of training data containing anomalies with negligibly affected performance. The shift to fully unsupervised training simplifies dataset creation and broadens possible applications.

In this paper, we propose a novel task IntentQA, a special VideoQA task focusing on video intent reasoning, which has become increasingly important for AI with its advantages in equipping AI agents with the capability of reasoning beyond mere recognition in daily tasks. We also contribute a large-scale VideoQA dataset for this task. We propose a Context-aware Video Intent Reasoning model (CaVIR) consisting of i) Video Query Language (VQL) for better cross-modal representation of the situational context, ii) Contrastive Learning module for utilizing the contrastive context, and iii) Commonsense Reasoning module for incorporating the commonsense context. Comprehensive experiments on this challenging task demonstrate the effectiveness of each model component, the superiority of our full model over other baselines, and the generalizability of our model to a new VideoQA task. The dataset and codes are open-sourced at: https://github.com/JoseponLee/IntentQA.git

Tracking any given object(s) spatially and temporally is a common purpose in Visual Object Tracking (VOT) and Video Object Segmentation (VOS). Joint tracking and segmentation have been attempted in some studies but they often lack full compatibility of both box and mask in initialization and prediction, and mainly focus on single-object scenarios. To address these limitations, this paper proposes a Multi-object Mask-box Integrated framework for unified Tracking and Segmentation, dubbed MITS. Firstly, the unified identification module is proposed to support both box and mask reference for initialization, where detailed object information is inferred from boxes or directly retained from masks. Additionally, a novel pinpoint box predictor is proposed for accurate multi-object box prediction, facilitating target-oriented representation learning. All target objects are processed simultaneously from encoding to propagation and decoding, as a unified pipeline for VOT and VOS. Experimental results show MITS achieves state-of-the-art performance on both VOT and VOS benchmarks. Notably, MITS surpasses the best prior VOT competitor by around 6% on the GOT-10k test set, and significantly improves the performance of box initialization on VOS benchmarks. The code is available at https://github.com/yoxu515/MITS.

Modern object detectors have taken the advantages of backbone networks pre-trained on large scale datasets. Except for the backbone networks, however, other components such as the detector head and the feature pyramid network (FPN) remain trained from scratch, which hinders the generalization capacity of detectors. In this study, we propose to integrally migrate pre-trained transformer encoder-decoders (imTED) to a detector, constructing a feature extraction path which is "fully pre-trained" so that detectors' generalization capacity is maximized. The essential differences between imTED with the baseline detector are twofold: (1) migrating the pre-trained transformer decoder to the detector head while removing the randomly initialized FPN from the feature extraction path; and (2) defining a multi-scale feature modulator (MFM) to enhance scale adaptability. Such designs not only reduce randomly initialized parameters significantly but also unify detector training with representation learning intendedly. Experiments on the MS COCO object detection dataset show that imTED consistently outperforms its counterparts by 2.4 AP. Without bells and whistles, imTED improves the state-of-the-art of few-shot object detection by up to 7.6 AP. Code is released at https://github.com/LiewFeng/imTED.

We propose a method for editing NeRF scenes with text-instructions. Given a NeRF of a scene and the collection of images used to reconstruct it, our method uses an image-conditioned diffusion model (InstructPix2Pix) to iteratively edit the input images while optimizing the underlying scene, resulting in an optimized 3D scene that respects the edit instruction. We demonstrate that our proposed method is able to edit large-scale, real-world scenes, and is able to accomplish more realistic, targeted edits than prior work.

Pre-trained point cloud models have found extensive applications in 3D understanding tasks like object classification and part segmentation. However, the prevailing strategy of full fine-tuning in downstream tasks leads to large per-task storage overhead for model parameters, which limits the efficiency when applying large-scale pre-trained models. Inspired by the recent success of visual prompt tuning (VPT), this paper attempts to explore prompt tuning on pre-trained point cloud models, to pursue an elegant balance between performance and parameter efficiency. We find while instance-agnostic static prompting, e.g. VPT, shows some efficacy in downstream transfer, it is vulnerable to the distribution diversity caused by various types of noises in real-world point cloud data. To conquer this limitation, we propose a novel Instance-aware Dynamic Prompt Tuning (IDPT) strategy for pre-trained point cloud models. The essence of IDPT is to develop a dynamic prompt generation module to perceive semantic prior features of each point cloud instance and generate adaptive prompt tokens to enhance the model's robustness. Notably, extensive experiments demonstrate that IDPT outperforms full fine-tuning in most tasks with a mere 7% of the trainable parameters, providing a promising solution to parameter-efficient learning for pre-trained point cloud models. Code is available at https://github.com/zyh16143998882/ICCV23-IDPT.

Continual Learning (CL) is the constant development of complex behaviors by building upon previously acquired skills. Yet, current CL algorithms tend to incur class-level forgetting as the label information is often quickly overwritten by new knowledge. This motivates attempts to mine instance-level discrimination by resorting to recent self-supervised learning (SSL) techniques. However, previous works have pointed that the self-supervised learning objective is essentially a trade-off between invariance to distortion and preserving sample information, which seriously hinders the unleashing of instance-level discrimination. In this work, we reformulate SSL from the information-theoretic perspective by disentangling the goal of instance-level discrimination, and tackle the trade-off to promote compact representations with maximally preserved invariance to distortion. On this basis, we develop a novel alternate learning paradigm to enjoy the complementary merits of instance-level and category-level supervision, which yields improved robustness against forgetting and better adaptation to each task. To verify the proposed method, we conduct extensive experiments on four different benchmarks using both class-incremental and task-incremental settings, where the leap in performance and thorough ablation studies demonstrate the efficacy and efficiency of our modeling strategy.

This paper presents one of the first learning-based NeRF 3D instance segmentation pipelines, dubbed as Instance Neural Radiance Field, or Instance-NeRF. Taking a NeRF pretrained from multi-view RGB images as input, Instance-NeRF can learn 3D instance segmentation of a given scene, represented as an instance field component of the NeRF model. To this end, we adopt a 3D proposal-based mask prediction network on the sampled volumetric features from NeRF, which generates discrete 3D instance masks. The coarse 3D mask prediction is then projected to image space to match 2D segmentation masks from different views generated by existing panoptic segmentation models, which are used to supervise the training of the instance field. Notably, beyond generating consistent 2D segmentation maps from novel views, Instance-NeRF can query instance information at any 3D point, which greatly enhances NeRF object segmentation and manipulation. Our method is also one of the first to achieve such results in pure inference. Experimented on synthetic and real-world NeRF datasets with complex indoor scenes, Instance-NeRF surpasses previous NeRF segmentation works and competitive 2D segmentation methods in segmentation performance on unseen views. Code and data are available at https://github.com/lyclyc52/Instance_NeRF.

Recent progress in generative models has resulted in models that produce both realistic as well as relevant images for most textual inputs. These models are being used to generate millions of images everyday, and hold the potential to drastically impact areas such as generative art, digital marketing and data augmentation. Given their outsized impact, it is important to ensure that the generated content reflects the artifacts and surroundings across the globe, rather than over-representing certain parts of the world. In this paper, we measure the geographical representativeness of common nouns (e.g., a house) generated through DALL.E 2 and Stable Diffusion models using a crowdsourced study comprising 540 participants across 27 countries. For deliberately underspecified inputs without country names, the generated images most reflect the surroundings of the United States followed by India, and the top generations rarely reflect surroundings from all other countries (average score less than 3 out of 5). Specifying the country names in the input increases the representativeness by 1.44 points on average on a 5-point Likert scale for DALL.E 2 and 0.75 for Stable Diffusion, however, the overall scores for many countries still remain low, highlighting the need for future models to be more geographically inclusive. Lastly, we examine the feasibility of quantifying the geographical representativeness of generated images without conducting user studies.

Fisheye image rectification is hindered by synthetic models producing poor results for real-world correction. To address this, we propose a Dual Diffusion Architecture (DDA) for fisheye rectification that offers better practicality. The DDA leverages Denoising Diffusion Probabilistic Models (DDPMs) to gradually introduce bidirectional noise, allowing the synthesized and real images to develop into a consistent noise distribution. As a result, our network can perceive the distribution of unlabelled real fisheye images without relying on a transfer network, thus improving the performance of real fisheye correction. Additionally, we design an unsupervised one-pass network that generates a plausible new condition to strengthen guidance and address the non-negligible indeterminacy between the prior condition and the target. It can significantly affect the rectification task, especially in cases where radial distortion causes significant artifacts. This network can be regarded as an alternate scheme for fast producing reliable results without iterative inference. Compared to the state-of-the-art methods, our approach achieves superior performance in both synthetic and real fisheye image corrections.

Spiking Neural Networks (SNNs) are well known as a promising energy-efficient alternative to conventional artificial neural networks. Subject to the preconceived impression that SNNs are sparse firing, the analysis and optimization of inherent redundancy in SNNs have been largely overlooked, thus the potential advantages of spike-based neuromorphic computing in accuracy and energy efficiency are interfered. In this work, we pose and focus on three key questions regarding the inherent redundancy in SNNs. We argue that the redundancy is induced by the spatio-temporal invariance of SNNs, which enhances the efficiency of parameter utilization but also invites lots of noise spikes. Further, we analyze the effect of spatio-temporal invariance on the spatio-temporal dynamics and spike firing of SNNs. Then, motivated by these analyses, we propose an Advance Spatial Attention (ASA) module to harness SNNs' redundancy, which can adaptively optimize their membrane potential distribution by a pair of individual spatial attention sub-modules. In this way, noise spike features are accurately regulated. Experimental results demonstrate that the proposed method can significantly drop the spike firing with better performance than state-of-the-art baselines. Our code is available in https://github.com/BICLab/ASA-SNN.

Contemporary domain adaptation offers a practical solution for achieving cross-domain transfer of semantic segmentation between labelled source data and unlabeled target data. These solutions have gained significant popularity; however, they require the model to be retrained when the test environment changes. This can result in unbearable costs in certain applications due to the time-consuming training process and concerns regarding data privacy. One-shot domain adaptation methods attempt to overcome these challenges by transferring the pre-trained source model to the target domain using only one target data. Despite this, the referring style transfer module still faces issues with computation cost and over-fitting problems. To address this problem, we propose a novel framework called Informative Data Mining (IDM) that enables efficient one-shot domain adaptation for semantic segmentation. Specifically, IDM provides an uncertainty-based selection criterion to identify the most informative samples, which facilitates quick adaptation and reduces redundant training. We then perform a model adaptation method using these selected samples, which includes patch-wise mixing and prototype-based information maximization to update the model. This approach effectively enhances adaptation and mitigates the overfitting problem. In general, we provide empirical evidence of the effectiveness and efficiency of IDM. Our approach outperforms existing methods and achieves a new state-of-the-art one-shot performance of 56.7%/55.4% on the GTA5/SYNTHIA to Cityscapes adaptation tasks, respectively. The code will be released at https://github.com/yxiwang/IDM.

Toward infinite-scale 3D city synthesis, we propose a novel framework, InfiniCity, which constructs and renders an unconstrainedly large and 3D-grounded environment from random noises. InfiniCity decomposes the seemingly impractical task into three feasible modules, taking advantage of both 2D and 3D data. First, an infinite-pixel image synthesis module generates arbitrary-scale 2D maps from the bird's-eye view. Next, an octree-based voxel completion module lifts the generated 2D map to 3D octrees. Finally, a voxel-based neural rendering module texturizes the voxels and renders 2D images. InfiniCity can thus synthesize arbitrary-scale and traversable 3D city environments. We quantitatively and qualitatively demonstrate the efficacy of the proposed framework.

There is a recently discovered and intriguing phenomenon called Neural Collapse: at the terminal phase of training a deep neural network for classification, the within-class penultimate feature means and the associated classifier vectors of all flat classes collapse to the vertices of a simplex Equiangular Tight Frame (ETF). Recent work has tried to exploit this phenomenon by fixing the related classifier weights to a pre-computed ETF to induce neural collapse and maximize the separation of the learned features when training with imbalanced data. In this work, we propose to fix the linear classifier of a deep neural network to a Hierarchy-Aware Frame (HAFrame), instead of an ETF, and use a cosine similarity-based auxiliary loss to learn hierarchy-aware penultimate features that collapse to the HAFrame. We demonstrate that our approach reduces the mistake severity of the model's predictions while maintaining its top-1 accuracy on several datasets of varying scales with hierarchies of heights ranging from 3 to 12. Code: https://github.com/ltong1130ztr/HAFrame.

In recent years, depth recovery based on deep networks has achieved great success. However, the existing state-of-the-art network designs perform like black boxes in depth recovery tasks, lacking a clear mechanism. Utilizing the property that there is a large amount of non-local common characteristics in depth images, we propose a novel model-guided depth recovery method, namely the DC-NLAR model. A non-local auto-regressive regular term is also embedded into our model to capture more non-local depth information. To fully use the excellent performance of neural networks, we develop a deep image prior to better describe the characteristic of depth images. We also introduce an implicit data consistency term to tackle the degenerate operator with high heterogeneity. We then unfold the proposed model into networks by using the half-quadratic splitting algorithm. This proposed method is experimented on the NYU-Depth V2 and SUN RGB-D datasets, and the experimental results achieve comparable performance to that of deep learning methods.

We explore the problem of Incremental Generalized Category Discovery (IGCD). This is a challenging category-incremental learning setting where the goal is to develop models that can correctly categorize images from previously seen categories, in addition to discovering novel ones. Learning is performed over a series of time steps where the model obtains new labeled and unlabeled data, and discards old data, at each iteration. The difficulty of the problem is compounded in our generalized setting as the unlabeled data can contain images from categories that may or may not have been observed before. We present a new model for IGCD which combines non-parametric categorization with efficient image sampling to mitigate catastrophic forgetting. To quantify performance, we propose a new benchmark dataset named iNatIGCD that is motivated by a real-world fine-grained visual categorization task. In our experiments we outperform existing related methods.

Large-scale noisy web image-text datasets have been proven to be efficient for learning robust vision-language models. However, to transfer them to the task of video retrieval, models still need to be fine-tuned on hand-curated paired text-video data to adapt to the diverse styles of video descriptions. To address this problem without the need for hand-annotated pairs, we propose a new setting, text-video retrieval with uncurated & unpaired data, that uses only text queries together with uncurated web videos during training without any paired text-video data. To this end, we propose an approach, In-Style, that learns the style of the text queries and transfers it to uncurated web videos. Moreover, to improve generalization, we show that one model can be trained with multiple text styles. To this end, we introduce a multi-style contrastive training procedure, that improves the generalizability over several datasets simultaneously. We evaluate our model on retrieval performance over multiple datasets to demonstrate the advantages of our style transfer framework on the new task of uncurated & unpaired text-video retrieval and improve state-of-the-art performance on zero-shot text-video retrieval.

Programs offer compactness and structure that makes them an attractive representation for visual data. We explore how code rewriting can be used to improve systems for inferring programs from visual data. We first propose Sparse Intermittent Rewrite Injection (SIRI), a framework for unsupervised bootstrapped learning. SIRI sparsely applies code rewrite operations over a dataset of training programs, injecting the improved programs back into the training set. We design a family of rewriters for visual programming domains: parameter optimization, code pruning, and code grafting. For three shape programming languages in 2D and 3D, we experimentally validate that using SIRI with our family of rewriters improves performance: better reconstructions and faster convergence rates, compared with bootstrapped learning methods that do not use rewriters or use them naively. Finally, we demonstrate that our family of rewriters can be effectively employed at test time to improve the output of SIRI predictions. For 2D and 3D CSG, we outperform or match the reconstruction performance of recent domain-specific neural architectures, while producing more parsimonious programs, that use significantly fewer primitives.