Most existing object tracking research focuses on pedestrians and autonomous driving while ignoring sports scenes. When general object tracking models are used for ball tracking, there are often problems, such as detection omissions due to small object sizes and trajectory loss due to occlusion. To address these challenges, we propose a ball detection and tracking model called HMMATrack based on multiscale feature enhancement and multilevel collaborative matching to improve ball-tracking results from the entire process of sampling, feature extraction, detection, and tracking. It includes a Heuristic Compound Sampling Strategy to deal with tiny sizes and imbalanced data samples; an MNet-based detection module to improve the ball detection accuracy; and a multilevel cooperative matching and automatic trajectory correction tracking algorithm that can quickly and accurately correct the ball’s trajectory. We also hand-annotated SportsTrack, a ball-tracking dataset containing soccer, basketball, and volleyball scenes. Extensive experiments are conducted on the SportsTrack, demonstrating that our proposed HMMATrack model outperforms other representative state-of-the-art models in ball detection and tracking.

Most of the existing low-light enhancement research focuses on global illumination enhancement while ignoring the issues of brightness unevenness and color distortion. To address this dilemma, we propose a low-light image enhancement method that can achieve good performance in luminance adjustment and color correction simultaneously. Specifically, the Luminance Adjustment Module is designed to model the global luminance adjustment parameters while taking into account the relationship between global and local illumination features, in order to prevent overexposure or underexposure. Furthermore, we design a Color Correction Module based on the attention mechanism, which utilizes the attention mechanism to capture global color features and correct the color deviation in the illumination-enhanced image. Additionally, we design a color loss function based on a 14-dimensional statistical feature vector related to color, enabling further restoration of the image’s true color. We conduct empirical studies on multiple public low-light datasets, demonstrating that the proposed method outperforms other representative state-of-the-art models regarding illumination enhancement and color correction.

Sports competition is one of the most popular programs for many audiences. Tracking the players in sports game videos from broadcasts is a nontrivial challenge for computer vision researchers. In sports videos, the direction of an athlete’s movement changes quickly and unpredictably. Mutual occlusion between athletes is also more frequent in team competitions. However, the rich temporal contexts among the adjacent frames have been excluded from consideration. To address this dilemma, we propose an online transformer‐based learnable framework in an end‐to‐end fashion. We use a transformer architecture to extract the temporal contexts between the successive frames and add them to the network training, which is robust to occlusion and complex direction changes in multiplayer tracking. We demonstrate the effectiveness of our method on three sports video datasets by comparing them with recently advanced multiplayer trackers.

Neural language models have achieved remarkable progress in semantic representation learning. However, cross-domain representation learning still suffers from prominent semantic noise propagation issues. Existing methods still face challenges in cross-domain semantic modeling, including limited robustness across different semantic granularities, difficulty in separating transferable semantics from task-irrelevant semantic interference, and insufficient adaptability to specialized scenarios. These issues may reduce feature discriminability in fine-grained semantic tasks and complex application settings. To address these problems, we propose the Diffusion-Coded Attention Network (DCANet), a novel cross-domain representation learning architecture with three synergistic core modules: a multi-granular parallel diffusion masking mechanism for cross-scale context fusion via stochastic path activation, an implicit semantic encoder that distills domain-invariant patterns into adaptive bias codes via shared latent manifolds, and a self-correcting attention topology realizing dynamic semantic purification via closed-loop interactions between local features and global bias states. Extensive evaluations are conducted on nine well-recognized benchmark datasets to verify DCANet’s effectiveness and reliability. Experimental results show that DCANet attains state-of-the-art results on the majority of the benchmark datasets, with significant accuracy improvements on text classification and sentiment analysis tasks.

Implicit hate speech is difficult to detect because hostile intent is often conveyed through metaphor, irony, coded expressions, stereotypes, or culturally situated allusions rather than direct insults. We first construct a fine-grained Chinese implicit hate speech dataset containing 19,939 samples from Zhihu and Baidu Tieba, covering nine target categories and three expression labels. Based on this dataset, we propose FICHS, a fine-grained and implicit Chinese hate speech detection framework with three synergistic modules built on a RoBERTa-based base detector: an Explicit–Implicit Supervised Contrastive Module that learns a discriminative semantic space between explicit and implicit hate speech, a Semantic Opacity-Aware Routing and Adjudication module that selectively routes low-confidence but high-implicitness samples to an LLM for secondary judgment, and a Pattern-Guided Retrieval-Augmented Recovery Module that retrieves socio-cultural risk patterns to support second-round review and false-negative recovery. Experimental results demonstrate that FICHS effectively improves Chinese implicit hate speech detection, achieving a weighted F1 score of 0.8391, with precision of 0.8428 and recall of 0.8396, outperforming both the SOTA detector and standalone LLM baselines.

Large Language Model (LLM)-driven social simulation offers a controllable approach for analyzing crisis responses, but existing work on online crises often emphasizes either user engagement prediction or opinion and eWOM evolution. This separation is insufficient for online firestorms, where crisis impact emerges from the coupling between what users express and how participation expands over time. In such events, responsibility attribution, KOL influence, electronic Word of Mouth (eWOM), and user engagement jointly shape collective reactions. To address this gap, we introduce the Crisis Response Interaction Simulation Pipeline (CRISP), a training-free LLM-driven multi-agent framework for online firestorm simulation. CRISP integrates an eWOM Perception module for responsibility attribution and attitude formation with an Engagement Mechanism for predicting participation under evolving KOL influence. Experiments on four heterogeneous Weibo online firestorms across beverage, automobile manufacturing, food service, and education domains show that CRISP reproduces major eWOM and engagement trajectories across different activity scales and interaction structures. Counterfactual interventions on emotional composition and responsibility attribution further produce directionally consistent responses, suggesting mechanism-level validity beyond trajectory fitting. These findings indicate that CRISP provides a framework for analyzing online firestorm evolution and evaluating crisis communication strategies in controllable simulation environments.