Facial affective behavior analysis (FABA) is crucial for understanding human mental states from images. However, traditional approaches primarily deploy models to discriminate among discrete emotion categories, and lack the fine granularity and reasoning capability for complex facial behaviors. The advent of Multi-modal Large Language Models (MLLMs) has been proven successful in general visual understanding tasks. However, directly harnessing MLLMs for FABA is challenging due to the scarcity of datasets and benchmarks, neglecting facial prior knowledge, and low training efficiency. To address these challenges, we introduce (i) an instruction-following dataset for two FABA tasks, e.g., emotion and action unit recognition, (ii) a benchmark FABA-Bench with a new metric considering both recognition and generation ability, and (iii) a new MLLM "EmoLA" as a strong baseline to the community. Our initiative on the dataset and benchmarks reveal the nature and rationale of facial affective behaviors, i.e., fine-grained facial movement, interpretability, and reasoning. Moreover, to build an effective and efficient FABA MLLM, we introduce a facial prior expert module with face structure knowledge and a low-rank adaptation module into pre-trained MLLM. We conduct extensive experiments on FABA-Bench and four commonly-used FABA datasets. The results demonstrate that the proposed facial prior expert can boost the performance and EmoLA achieves the best results on our FABA-Bench. On commonly-used FABA datasets, EmoLA is competitive rivaling task-specific state-of-the-art models.
Learning to build 3D scene graphs is essential for real-world perception in a structured and rich fashion. However, previous 3D scene graph generation methods utilize a fully supervised learning manner and require a large amount of entity-level annotation data of objects and relations, which is extremely resource-consuming and tedious to obtain. To tackle this problem, we propose 3D-VLAP, a weakly-supervised 3D scene graph generation method via Visual-Linguistic Assisted Pseudo-labeling. Specifically, our 3D-VLAP exploits the superior ability of current large-scale visual-linguistic models to align the semantics between texts and 2D images, as well as the naturally existing correspondences between 2D images and 3D point clouds, and thus implicitly constructs correspondences between texts and 3D point clouds. First, we establish the positional correspondence from 3D point clouds to 2D images via camera intrinsic and extrinsic parameters, thereby achieving alignment of 3D point clouds and 2D images. Subsequently, a large-scale cross-modal visual-linguistic model is employed to indirectly align 3D instances with the textual category labels of objects by matching 2D images with object category labels. The pseudo labels for objects and relations are then produced for 3D-VLAP model training by calculating the similarity between visual embeddings and textual category embeddings of objects and relations encoded by the visual-linguistic model, respectively. Ultimately, we design an edge self-attention based graph neural network to generate scene graphs of 3D point cloud scenes. Extensive experiments demonstrate that our 3D-VLAP achieves comparable results with current advanced fully supervised methods, meanwhile significantly alleviating the pressure of data annotation.
In this paper, we study the harmlessness alignment problem of multimodal large language models~(MLLMs). We conduct a systematic empirical analysis of the harmlessness performance of representative MLLMs and reveal that the image input poses the alignment vulnerability of MLLMs. Inspired by this, we propose a novel jailbreak method named HADES, which hides and amplifies the harmfulness of the malicious intent within the text input, using meticulously crafted images. Experimental results show that HADES can effectively jailbreak existing MLLMs, which achieves an average Attack Success Rate~(ASR) of 90.26% for LLaVA-1.5 and 71.60% for Gemini Pro Vision. Our code and data will be publicly released.
Due to their unprecedented ability to process and respond to various types of data, Multimodal Large Language Models (MLLMs) are constantly defining the new boundary of Artificial General Intelligence (AGI). As these advanced generative models increasingly form collaborative networks for complex tasks, the integrity and security of these systems are crucial. Our paper, ``The Wolf Within'', explores a novel vulnerability in MLLM societies - the indirect propagation of malicious content. Unlike direct harmful output generation for MLLMs, our research demonstrates how a single MLLM agent can be subtly influenced to generate prompts that, in turn, induce other MLLM agents in the society to output malicious content. This subtle, yet potent method of indirect influence marks a significant escalation in the security risks associated with MLLMs. Our findings reveal that, with minimal or even no access to MLLMs' parameters, an MLLM agent, when manipulated to produce specific prompts or instructions, can effectively ``infect'' other agents within a society of MLLMs. This infection leads to the generation and circulation of harmful outputs, such as dangerous instructions or misinformation, across the society. We also show the transferability of these indirectly generated prompts, highlighting their possibility in propagating malice through inter-agent communication. This research provides a critical insight into a new dimension of threat posed by MLLMs, where a single agent can act as a catalyst for widespread malevolent influence. Our work underscores the urgent need for developing robust mechanisms to detect and mitigate such covert manipulations within MLLM societies, ensuring their safe and ethical utilization in societal applications. Our implementation is released at \url{https://github.com/ChengshuaiZhao0/The-Wolf-Within.git}.
Current image manipulation primarily centers on static manipulation, such as replacing specific regions within an image or altering its overall style. In this paper, we introduce an innovative dynamic manipulation task, subject repositioning. This task involves relocating a user-specified subject to a desired position while preserving the image's fidelity. Our research reveals that the fundamental sub-tasks of subject repositioning, which include filling the void left by the repositioned subject, reconstructing obscured portions of the subject and blending the subject to be consistent with surrounding areas, can be effectively reformulated as a unified, prompt-guided inpainting task. Consequently, we can employ a single diffusion generative model to address these sub-tasks using various task prompts learned through our proposed task inversion technique. Additionally, we integrate pre-processing and post-processing techniques to further enhance the quality of subject repositioning. These elements together form our SEgment-gEnerate-and-bLEnd (SEELE) framework. To assess SEELE's effectiveness in subject repositioning, we assemble a real-world subject repositioning dataset called ReS. Our results on ReS demonstrate the quality of repositioned image generation.
RGBT tracking has been widely used in various fields such as robotics, surveillance processing, and autonomous driving. Existing RGBT trackers fully explore the spatial information between the template and the search region and locate the target based on the appearance matching results. However, these RGBT trackers have very limited exploitation of temporal information, either ignoring temporal information or exploiting it through online sampling and training. The former struggles to cope with the object state changes, while the latter neglects the correlation between spatial and temporal information. To alleviate these limitations, we propose a novel Temporal Adaptive RGBT Tracking framework, named as TATrack. TATrack has a spatio-temporal two-stream structure and captures temporal information by an online updated template, where the two-stream structure refers to the multi-modal feature extraction and cross-modal interaction for the initial template and the online update template respectively. TATrack contributes to comprehensively exploit spatio-temporal information and multi-modal information for target localization. In addition, we design a spatio-temporal interaction (STI) mechanism that bridges two branches and enables cross-modal interaction to span longer time scales. Extensive experiments on three popular RGBT tracking benchmarks show that our method achieves state-of-the-art performance, while running at real-time speed.
To satisfy the high-resolution requirements of direction-of-arrival (DOA) estimation, conventional deep neural network (DNN)-based methods using grid idea need to significantly increase the number of output classifications and also produce a huge high model complexity. To address this problem, a multi-level tree-based DNN model (TDNN) is proposed as an alternative, where each level takes small-scale multi-layer neural networks (MLNNs) as nodes to divide the target angular interval into multiple sub-intervals, and each output class is associated to a MLNN at the next level. Then the number of MLNNs is gradually increasing from the first level to the last level, and so increasing the depth of tree will dramatically raise the number of output classes to improve the estimation accuracy. More importantly, this network is extended to make a multi-emitter DOA estimation. Simulation results show that the proposed TDNN performs much better than conventional DNN and root-MUSIC at extremely low signal-to-noise ratio (SNR), and can achieve Cramer-Rao lower bound (CRLB). Additionally, in the multi-emitter scenario, the proposed Q-TDNN has also made a substantial performance enhancement over DNN and Root-MUSIC, and this gain grows as the number of emitters increases.
Graph Neural Networks (GNNs) have emerged as a powerful tool for representation learning on graphs, but they often suffer from overfitting and label noise issues, especially when the data is scarce or imbalanced. Different from the paradigm of previous methods that rely on single-node confidence, in this paper, we introduce a novel Class-wise Selection for Graph Neural Networks, dubbed CSGNN, which employs a neighbor-aggregated latent space to adaptively select reliable nodes across different classes. Specifically, 1) to tackle the class imbalance issue, we introduce a dynamic class-wise selection mechanism, leveraging the clustering technique to identify clean nodes based on the neighbor-aggregated confidences. In this way, our approach can avoid the pitfalls of biased sampling which is common with global threshold techniques. 2) To alleviate the problem of noisy labels, built on the concept of the memorization effect, CSGNN prioritizes learning from clean nodes before noisy ones, thereby iteratively enhancing model performance while mitigating label noise. Through extensive experiments, we demonstrate that CSGNN outperforms state-of-the-art methods in terms of both effectiveness and robustness.
Traditional deep neural networks (DNNs) have bad performance on estimating off-grid angles, and the most direct solution is to increase the number of output classes for improving angular resolution. But more output classes can weaken the model accuracy of DNNs and thus decreasing the direction-of-arrival (DOA) estimation accuracy. In this work, a tree-model based deep neural networks (TDNN) is proposed, which contains H layers and each layer is consist of multiple small-scale DNNs. From the first layer to the last layer of TDNN, the angular region is gradually divided into smaller subregions by these DNNs, and the estimated DOA is finally obtained by cumulative calculating the classification results of all the layers. TDNN can improve the angular resolution by increasing the number of layers or the number of DNNs in any layer instead of changing the structure of single DNN, so the model accuracy of TDNN will not decrease with the improvement of angular resolution and its estimation performance is also stable. In addition, the Q-TDNN method is also proposed for multi-sources DOA estimation, which can obtain Q different DOAs from the same signals by combining Q independent and parallel TDNNs. The simulation results validate TDNN has much better estimation performance than traditional methods in both single-source and multi-sources cases, especially at low signal-to-noise ratio (SNR).
Enabling robots to understand language instructions and react accordingly to visual perception has been a long-standing goal in the robotics research community. Achieving this goal requires cutting-edge advances in natural language processing, computer vision, and robotics engineering. Thus, this paper mainly investigates the potential of integrating the most recent Large Language Models (LLMs) and existing visual grounding and robotic grasping system to enhance the effectiveness of the human-robot interaction. We introduce the WALL-E (Embodied Robotic WAiter load lifting with Large Language model) as an example of this integration. The system utilizes the LLM of ChatGPT to summarize the preference object of the users as a target instruction via the multi-round interactive dialogue. The target instruction is then forwarded to a visual grounding system for object pose and size estimation, following which the robot grasps the object accordingly. We deploy this LLM-empowered system on the physical robot to provide a more user-friendly interface for the instruction-guided grasping task. The further experimental results on various real-world scenarios demonstrated the feasibility and efficacy of our proposed framework. See the project website at: https://star-uu-wang.github.io/WALL-E/