ABSTRACT Image‐based person search, which locates and identifies a query person from uncropped gallery images, remains challenging due to occlusions, viewpoint changes, and ambiguous appearances. While existing methods increasingly leverage visual context, they are fundamentally limited by static fusion strategies for multi‐view features and their neglect of semantic context from other modalities. To overcome these limitations, we propose the dynamic cross‐modal context fusion (DCMCF) framework, which establishes a core multi‐modal fusion architecture by explicitly leveraging self‐generated textual descriptions to provide semantics‐driven matching guidance. DCMCF consists of three core components: (1) multi‐modal feature extraction, which employs an enhanced multi‐view visual encoder with dynamic gating and a zero‐shot pipeline using pre‐trained VLMs to generate multi‐level textual descriptions from images; (2) a dynamic cross‐modal fusion (DCMF) module, which hierarchically and adaptively fuses the visual and textual features through cross‐modal attention and input‐dependent gating; and (3) a text‐guided inter‐image group context ranking (T‐IGCR) algorithm, which refines retrieval results by measuring holistic image consistency in both visual and textual spaces. Experiments on CUHK‐SYSU and PRW datasets demonstrate state‐of‐the‐art performance, achieving 96.9%/97.3% (mAP/top‐1) and 56.1%/91.2% , respectively. This work demonstrates the significant potential of dynamic cross‐modal context fusion for advancing image‐based person search.

ABSTRACT In scenarios with heavy smog, the accuracy of object detection in low‐light visible polarization images significantly decreases. To address this issue, we propose a dual‐branch enhancement and multi‐modal fusion network for object detection in low‐light visible polarization images in dense smog environments. Specifically, the network consists of an image enhancement stage and an object detection stage. In the image enhancement stage, a dual‐branch enhancement structure comprising greyscale feature map prediction and atmospheric light transmission network is proposed to remove noise from the images and enhance texture information, jointly generating enhanced visible polarization images. In the object detection stage, feature maps of the enhanced visible polarization images and the degree of visible polarization images are fused, and their fused texture‐enhanced feature maps are fed into the detection module for object detection. Additionally, we have collected a dataset of low‐light visible polarization images under real smog conditions. Extensive experiments demonstrate that our method can generate visually improved enhanced images and significantly increase detection accuracy and the number of detected objects in low‐light and dense smog environments.

The main objective of multiagent reinforcement learning is to achieve a global optimal policy. It is difficult to evaluate the value function with high-dimensional state space. Therefore, we transfer the problem of multiagent reinforcement learning into a distributed optimization problem with constraint terms. In this problem, all agents share the space of states and actions, but each agent only obtains its own local reward. Then, we propose a distributed optimization with fractional order dynamics to solve this problem. Moreover, we prove the convergence of the proposed algorithm and illustrate its effectiveness with a numerical example.

Knowledge representation learning (KRL) encodes enormous structured information with entities and relations into a continuous low-dimensional semantic space. Most conventional methods solely focus on learning knowledge representation from single modality, yet neglect the complementary information from others. The more and more rich available multi-modal data on Internet also drive us to explore a novel approach for KRL in multi-modal way, and overcome the limitations of previous single-modal based methods. This paper proposes a novel multi-modal knowledge representation learning (MM-KRL) framework which attempts to handle knowledge from both textual and visual modal web data. It consists of two stages, i.e., webly-supervised multi-modal relationship mining, and bi-enhanced cross-modal knowledge representation learning. Compared with existing knowledge representation methods, our framework has several advantages: (1) It can effectively mine multi-modal knowledge with structured textual and visual relationships from web automatically. (2) It is able to learn a common knowledge space which is independent to both task and modality by the proposed Bi-enhanced Cross-modal Deep Neural Network (BC-DNN). (3) It has the ability to represent unseen multi-modal relationships by transferring the learned knowledge with isolated seen entities and relations into unseen relationships. We build a large-scale multi-modal relationship dataset (MMR-D) and the experimental results show that our framework achieves excellent performance in zero-shot multi-modal retrieval and visual relationship recognition.