ACM MM 2024 Papers

Maintaining temporal stability is crucial in multi-agent trajectory prediction. Insufficient regularization to uphold this temporal stability often results in fluctuations in kinematic states, leading to inconsistent predictions and the amplification of errors. In this study, we introduce a framework called Multi-Agent Trajectory prediction via neural interaction Energy (MATE). This framework assesses the interactive motion of agents by employing neural interaction energy, which captures the dynamics of interactions and illustrates their influence on the future trajectories of agents. To bolster temporal stability, we introduce two constraints: inter-agent interaction constraint and intra-agent motion constraint. These constraints work together to ensure temporal stability at both the system and agent levels, effectively mitigating prediction fluctuations inherent in multi-agent systems. Comparative evaluations against previous methods on four diverse datasets, including simulated and real-world scenarios, highlight the superior prediction accuracy and generalization capabilities of our model.

Fake audio detection is an emerging active topic. A growing number of literatures have aimed to detect fake utterance, which are mostly generated by Text-to-speech (TTS) or voice conversion (VC). However, countermeasures against impersonation remain an underexplored area. Impersonation is a fake type that involves an imitator replicating specific traits and speech style of a target speaker. Unlike TTS and VC, which often leave digital traces or signal artifacts, impersonation involves live human beings producing entirely natural speech, rendering the detection of impersonation audio a challenging task. Thus, we propose a novel method that integrates speaker profiles into the process of impersonation audio detection. Speaker profiles are inherent characteristics that are challenging for impersonators to mimic accurately, such as speaker's age, job. We aim to leverage these features to extract discriminative information for detecting impersonation audio. Moreover, there is no large impersonated speech corpora available for quantitative study of impersonation impacts. To address this gap, we further design the first large-scale, diverse-speaker Chinese impersonation dataset, named ImPersonation Audio Detection (IPAD), to advance the community's research on impersonation audio detection. We evaluate several existing fake audio detection methods on our proposed dataset IPAD, demonstrating its necessity and the challenges. Additionally, our findings reveal that incorporating speaker profiles can significantly enhance the model's performance in detecting impersonation audio.

Recent years have witnessed remarkable progress in the development of large vision-language models (LVLMs). Benefiting from the strong language backbones and efficient cross-modal alignment strategies, LVLMs exhibit surprising capabilities to perceive visual signals and perform visually grounded reasoning. However, the capabilities of LVLMs have not been comprehensively and quantitatively evaluated. Most existing multi-modal benchmarks require task-oriented input-output formats, posing great challenges to automatically assess the free-form text output of LVLMs. To effectively leverage the annotations available and reduce the manual efforts required for constructing new benchmarks, we propose to re-formulate existing benchmarks into unified LVLM-compatible formats. Through systematic data collection and reformulation, we present ReForm-Eval benchmark, offering substantial data for evaluating various capabilities of LVLMs. Through extensive experiments and analysis in ReForm-Eval, we demonstrate the comprehensiveness and reliability of ReForm-Eval in assessing various LVLMs. Our benchmark and evaluation framework is now available at https://github.com/FudanDISC/ReForm-Eval

Text-to-image (T2I) diffusion models enjoy great popularity and many individuals and companies build their applications based on publicly released T2I diffusion models. Previous studies have demonstrated that backdoor attacks can elicit T2I diffusion models to generate unsafe target images through textual triggers. However, existing backdoor attacks typically demand substantial tuning data for poisoning, limiting their practicality and potentially degrading the overall performance of T2I diffusion models. To address these issues, we propose EvilEdit, a training-free and data-free backdoor attack against T2I diffusion models. EvilEdit directly edits the projection matrices in the cross-attention layers to achieve projection alignment between a trigger and the corresponding backdoor target. We preserve the functionality of the backdoored model using a protected whitelist to ensure the semantic of non-trigger words is not accidentally altered by the backdoor. We also propose a visual target attack EvilEdit VTA, enabling adversaries to use specific images as backdoor targets. We conduct empirical experiments on Stable Diffusion and the results demonstrate that the EvilEdit can backdoor T2I diffusion models within one second with up to 100% success rate. Furthermore, our EvilEdit modifies only 2.2% of the parameters and maintains the model's performance on benign prompts. Our code is available at https://github.com/haowang-cqu/EvilEdit.

Out-of-Distribution (OOD) detection is crucial when deploying machine learning models in open-world applications. The core challenge in OOD detection is mitigating the model's overconfidence on OOD data. While recent methods using auxiliary outlier datasets or synthesizing outlier features have shown promising OOD detection performance, they are limited due to costly data collection or simplified assumptions. In this paper, we propose a novel OOD detection framework FodFoM that innovatively combines multiple foundation models to generate two types of challenging fake outlier images for classifier training. The first type is based on BLIP-2's image captioning capability, CLIP's vision-language knowledge, and Stable Diffusion's image generation ability. Jointly utilizing these foundation models constructs fake outlier images which are semantically similar to but different from in-distribution (ID) images. For the second type, GroundingDINO's object detection ability is utilized to help construct pure background images by blurring foreground ID objects in ID images. The proposed framework can be flexibly combined with multiple existing OOD detection methods. Extensive empirical evaluations show that image classifiers with the help of constructed fake images can more accurately differentiate real OOD image from ID ones. New state-of-the-art OOD detection performance is achieved on multiple benchmarks. The code is available at https://github.com/Cverchen/ACMMM2024-FodFoM.

Universal object detectors aim to detect any object in any scene without human annotation, exhibiting superior generalization. However, the current universal object detectors show degraded performance in harsh weather, and their insufficient real-time capabilities limit their application. In this paper, we present Uni-YOLO, a universal detector designed for complex scenes with real-time performance. Uni-YOLO is a one-stage object detector that uses general object confidence to distinguish between objects and backgrounds, and employs a grid cell regression method for real-time detection. To improve its robustness in harsh weather conditions, the input of Uni-YOLO is adaptively enhanced with a physical model-based enhancement module. During training and inference, Uni-YOLO is guided by the extensive knowledge of the vision-language model CLIP. An object augmentation method is proposed to improve generalization in training by utilizing multiple source datasets with heterogeneous annotations. Furthermore, an online self-enhancement method is proposed to allow Uni-YOLO to further focus on specific objects through self-supervised fine-tuning in a given scene. Extensive experiments on public benchmarks and a UAV deployment are conducted to validate its superiority and practical value.

Currently, the information processing in a spatial domain alone has intrinsic limitations that hinder the deep network's effectiveness (performance) improvement in a single image deraining. Moreover, the deraining networks' structures and learning processes are becoming increasingly intricate, leading to challenges in structural lightweight, and training and testing efficiency. We propose a lightweight multi-domain multi-attention progressive network (M2PN) to handle these challenges. For performance improvement, the M2PN backbone applies a simple progressive CNN-based structure consisting of the S same recursive M2PN modules. This recursive backbone with a skip connection mechanism allows for better gradient flow and helps to effectively capture low-to-high-level/scales spatial features in progressive structure to improve contextual information acquisition. To further complement acquired spatial information for better deraining, we conduct spectral analysis on the frequency energy distribution of rain steaks, and theoretically present the relationship between the spectral bandwidths and the unique falling characteristics and special morphology of rain steaks. We present the frequency-channel attention (FcA) mechanism and the spatial-channel attention (ScA) mechanism to fuse frequency-channel features and spatial features better to distinguish and remove rain steaks. The simple recursive network structure and effective multi-domain multi-attention mechanism serve as the M2PN to achieve superior performance and facilitate fast convergence during training. Furthermore, the M2PN structure, with a small network component quantity, shallow network channels, and few convolutional kernels, requires only 168K parameters, which is 1 to 2 orders of magnitude lower than the existing SOTA networks. The experimental results demonstrate that even with such a few network parameters, M2PN still achieves the best overall performance.

Knowledge distillation (KD) is an established paradigm for transferring privileged knowledge from a cumbersome model to a more lightweight and efficient one. In recent years, logit-based KD methods are quickly catching up in performance with their feature-based counterparts. However, previous research has pointed out that logit-based methods are still fundamentally limited by two major issues in their training process, namely overconfident teacher and confirmation bias. Inspired by the success of cross-view learning in fields such as semi-supervised learning, in this work we introduce within-view and cross-view regularisations to standard logit-based distillation frameworks to combat the above cruxes. We also perform confidence-based soft label mining to improve the quality of distilling signals from the teacher, which further mitigates the confirmation bias problem. Despite its apparent simplicity, the proposed Consistency-Regularisation-based Logit Distillation (CRLD) significantly boosts student learning, setting new state-of-the-art results on the standard CIFAR-100, Tiny-ImageNet, and ImageNet datasets across a diversity of teacher and student architectures, whilst introducing no extra network parameters. Orthogonal to on-going logit-based distillation research, our method enjoys excellent generalisation properties and, without bells and whistles, boosts the performance of various existing approaches by considerable margins.

Missed polyps are the major risk factor for colorectal cancer. To minimize misdiagnosis, many methods have been developed. However, they either rely on laborious instance-level annotations, require labeling of prompt points, or lack the ability to filter noise proposals and detect polyps integrally, resulting in severe challenges in this area. In this paper, we propose a novel Cooperation-Based network (CBNet), a two-stage polyp detection framework supervised by image labels that removes wrong proposals through classification in collaboration with segmentation and obtains a more accurate detector by aggregating adaptive multi-level regional features. Specifically, we conduct a Cooperation-Based Region Proposal Network (CBRPN) to reduce the negative impact of noises by deleting proposals without polyps, enabling our network to capture polyp features. Moreover, to enhance location integrity and classification precision of polyps, we aggregate multi-level region of interest (ROI) features under the guidance of the backbone classification layer, namely Adaptive ROI Fusion Module (ARFM). Extensive experiments on the public and private datasets show that our method achieves state-of-the-art performance for weakly supervised methods and even outperforms full supervision in some terms. All code is available at https://github.com/dxqllp/CBNet.

Point tracking is a challenging task in computer vision, aiming to establish point-wise correspondence across long video sequences. Recent advancements have primarily focused on temporal modeling techniques to improve local feature similarity, often overlooking the valuable semantic consistency inherent in tracked points. In this paper, we introduce a novel approach leveraging language embeddings to enhance the coherence of frame-wise visual features related to the same object. Our proposed method, termed autogenic language embedding for visual feature enhancement, strengthens point correspondence in long-term sequences. Unlike existing visual-language schemes, our approach learns text embeddings from visual features through a dedicated mapping network, enabling seamless adaptation to various tracking tasks without explicit text annotations. Additionally, we introduce a consistency decoder that efficiently integrates text tokens into visual features with minimal computational overhead. Through enhanced visual consistency, our approach significantly improves tracking trajectories in lengthy videos with substantial appearance variations. Extensive experiments on widely-used tracking benchmarks demonstrate the superior performance of our method, showcasing notable enhancements compared to trackers relying solely on visual cues. The code will be available at https://github.com/SkyeSong38/ALTrack.

In real-world photography, local motion blur often arises from the interplay between moving objects and stationary backgrounds during exposure. Existing deblurring methods face challenges in addressing local motion deblurring due to (i) the presence of arbitrary localized blurs and uncertain blur extents; (ii) the limited ability to accurately identify specific blurs resulting from ambiguous motion boundaries. These limitations often lead to suboptimal solutions when estimating blur maps and generating final deblurred images. To that end, we propose a novel method named Motion-Uncertainty-Guided Network (MUGNet), which harnesses a probabilistic representational model to explicitly address the intricacies stemming from motion uncertainties. Specifically, MUGNet consists of two key components, i.e., motion-uncertainty quantification (MUQ) module and motion-masked separable attention (M2SA) module, serving for complementary purposes. Concretely, MUQ aims to learn a conditional distribution for accurate and reliable blur map estimation, while the M2SA module is to enhance the representation of regions influenced by local motion blur and static background, which is achieved by promoting the establishment of extensive global interactions. We demonstrate the superiority of our MUGNet with extensive experiments. The code is publicly available at: https://github.com/zeyuxiao1997/MUGNet.

Referring Remote Sensing Image Segmentation (RRSIS) is a challenging task that aims to identify specific regions in aerial images that are relevant to given textual conditions. Existing methods tend to adopt the paradigm of implicit optimization, utilizing a framework consisting of early cross-modal feature fusion and a fixed convolutional kernel-based predictor, neglecting the inherent inter-domain gap and conducting class-agnostic predictions. In this paper, we rethink the issues with the implicit optimization paradigm and address the RRSIS task from a dual-alignment perspective. Specifically, we prepend the dedicated Dual Alignment Network (DANet), including an explicit alignment strategy and a reliable agent alignment module. The explicit alignment strategy effectively reduces domain discrepancies by narrowing the inter-domain affinity distribution. Meanwhile, the reliable agent alignment module aims to enhance the predictor's multi-modality awareness and alleviate the impact of deceptive noise interference. Extensive experiments on two remote sensing datasets demonstrate the effectiveness of our proposed DANet in achieving superior segmentation performance without introducing additional learnable parameters compared to state-of-the-art methods.

Zero-shot anomaly detection (ZSAD) methods detect anomalies without prior access to known normal or abnormal samples within target categories. Existing methods typically rely on pretrained multimodal models, computing similarities between manually crafted textual features representing ''normal'' or ''abnormal'' semantics and image patch features to detect anomalies. However, the generic descriptions of ''abnormal'' often fail to precisely match diverse types of anomalies across different object categories. Additionally, computing feature similarities for single patches struggles to pinpoint specific locations of anomalies with various sizes and scales. To address these issues, we propose a novel ZSAD method called FiLo, comprising two components: adaptively learned Fine-Grained Description (FG-Des) and position-enhanced High-Quality Localization (HQ-Loc). FG-Des introduces fine-grained anomaly descriptions for each category using Large Language Models (LLMs) and employs adaptively learned textual templates to enhance the accuracy and interpretability of anomaly detection. HQ-Loc, utilizing Grounding DINO for preliminary localization, position-enhanced text prompts, and Multi-scale Multi-shape Cross-modal Interaction (MMCI) module, facilitates more accurate localization of anomalies of different sizes and shapes. Experimental results on datasets like MVTec and VisA demonstrate that FiLo significantly improves the performance of ZSAD in both detection and localization, achieving state-of-the-art performance with an image-level AUC of 83.9% and a pixel-level AUC of 95.9% on the VisA dataset. Code is available at https://github.com/CASIA-IVA-Lab/FiLo.

Text-to-image (T2I) generation is a pivotal and core interest within the realm of AI content generation. Amid the swift advancements of both open-source (such as Stable Diffusion) and proprietary (for example, DALLE, MidJourney) T2I models, there is a notable absence of a comprehensive and robust quantitative framework for evaluating their output quality. Traditional methods of quality assessment overlook the textual prompts when judging images; meanwhile, the advent of large multi-modal models (LMMs) introduces the capability to incorporate text prompts in evaluations, yet the challenge of fine-tuning these models for precise T2I quality assessment remains unresolved. In our study, we introduce the T2I-Scorer, a novel two-stage training methodology aimed at fine-tuning LMMs for T2I evaluation. For the first stage, we collect 397K GPT-4V-labeled question-answer pairs related to T2I evaluation. Termed as T2I-ITD, the pseudo-labeled dataset is analyzed and examined by human, and used for instruction tuning to improve the LMM's low-level quality perception. The first stage model, T2I-Scorer-IT, has reached superior accuracy on T2I evaluation than all kinds of existing T2I metrics under zero-shot settings. For the second stage, we define an explicit multi-task training scheme to further align the LMM with human opinion scores, and the fine-tuned T2I-Scorer can reach state-of-the-art accuracy on both image quality and image-text alignment perspectives with significant improvements. We anticipate the proposed metrics can serve as a reliable metric to gauge the ability of T2I generation models in the future. We will make code, data, and weights publicly available.

Drone-based crowd tracking faces difficulties in accurately identifying and monitoring objects from an aerial perspective, largely due to their small size and close proximity to each other, which complicates both localization and tracking. To address these challenges, we present the Density-aware Tracking (DenseTrack) framework. DenseTrack capitalizes on crowd counting to precisely determine object locations, blending visual and motion cues to improve the tracking of small-scale objects. It specifically addresses the problem of cross-frame motion to enhance tracking accuracy and dependability. DenseTrack employs crowd density estimates as anchors for exact object localization within video frames. These estimates are merged with motion and position information from the tracking network, with motion offsets serving as key tracking cues. Moreover, DenseTrack enhances the ability to distinguish small-scale objects using insights from the visual-language model, integrating appearance with motion cues. The framework utilizes the Hungarian algorithm to ensure the accurate matching of individuals across frames. Demonstrated on DroneCrowd dataset, our approach exhibits superior performance, confirming its effectiveness in scenarios captured by drones. Our code will be available at: https://github.com/Zebrabeast/DenseTrack.

Multi-task dense prediction plays an important role in the field of computer vision and has an abundant array of applications. Its main purpose is to reduce the amount of network training parameters by sharing network parameters while using the correlation between tasks to improve overall performance. We propose a task-conditional network that handles one task at a time and shares most network parameters to achieve these goals. Inspired by adapter tuning, we propose an adapter module that focuses on both spatial- and channel-wise information to extract features from the frozen encoder backbone. This approach not only reduces the number of training parameters, but also saves training time and memory resources by attaching a parallel adapter pathway to the encoder. We additionally use learnable task prompts to model different tasks and use these prompts to adjust some parameters of adapters to fit the network to diverse tasks. These task-conditional adapters are also applied to the decoder, which enables the entire network to switch between various tasks, producing better task-specific features and achieving excellent performance. Extensive experiments on two challenging multi-task benchmarks, NYUD-v2 and PASCAL-Context, show that our approach achieves state-of-the-art performance with excellent parameter, time, and memory efficiency. The code is available at https://github.com/jfzleo/Task-Conditional-Adapter.

We introduce HmPEAR, a novel dataset crafted for advancing research in 3D Human Pose Estimation (3D HPE) and Human Action Recognition (HAR), with a primary focus on outdoor environments. This dataset offers a synchronized collection of imagery, LiDAR point clouds, 3D human poses, and action categories. In total, the dataset encompasses over 300,000 frames collected from 10 distinct scenes and 25 diverse subjects. Among these, 250,000 frames of data contain 3D human pose annotations captured using an advanced motion capture system and further optimized for accuracy. Furthermore, the dataset annotates 40 types of daily human actions, resulting in over 6,000 action clips. Through extensive experimentation, we have demonstrated the quality of HmPEAR and highlighted the challenges it presents to current methodologies. Additionally, we propose baselines leveraging sequential images and point clouds for 3D HPE and HAR, which underscore the mutual reinforcement between them, highlighting the potential for cross-task synergies. The dataset is available at http://www.lidarhumanmotion.net/hmpear.

Open-domain multi-modal dialogue system heavily relies on visual information to generate contextually relevant responses. The existing open-domain multi-modal dialog generation methods ignore the complementary relationship between multiple modalities, and are difficult to integrate with LLMs. To tackle these challenges, we introduce AutoGraph, an innovative method for constructing visual context graphs automatically. We aim to structure complex information and seamlessly integrate it with large language models (LLMs), aligning information from multiple modalities at both semantic and structural levels. Specifically, we fully connect the text graphs and scene graphs, and then trim unnecessary edges via LLMs to automatically construct a visual context graph. Next, we design several graph sampling grammar for the first time to convert graph structures into sequence which is suitable for LLMs. Finally, we propose a two-stage fine-tuning strategy to allow LLMs to understand graph sampling grammar and generate responses. We validate our proposed method on text-based LLMs, and visual-based LLMs, respectively. Experimental results show that our proposed method achieves state-of-the-art performance on multiple public datasets.