Hu Zhang

With the rise of embodied agents and indoor service robots, object detection has become a critical component supporting semantic mapping, path planning, and human–robot interaction. However, indoor scenes often face challenges such as severe occlusion, large-scale variations, small and densely packed objects, and complex textures, making existing methods struggle in terms of both robustness and accuracy. This paper proposes MDF-YOLO, a multi-domain fusion detection framework based on Hölder regularity guidance. In the backbone, neck, and feature recovery stages, the framework introduces the CrossGrid Memory Block, Hölder-Based Regularity Guidance–Hierarchical Context Aggregation module, and Frequency-Guided Residual Block, achieving complementary feature modeling across the state space, spatial domain, and frequency domain. In particular, the HG-HCA module uses the Hölder regularity map as a guiding signal to balance the dynamic equilibrium between the macro and micro paths, thus achieving adaptive coordination between global consistency and local discriminability. Experimental results show that MDF-YOLO significantly outperforms mainstream detectors in metrics such as mAP@0.5, mAP@0.75, and mAP@0.5:0.95, achieving values of 0.7158, 0.6117, and 0.5814, respectively, while maintaining near real-time inference efficiency in terms of FPS and latency. Ablation studies further validate the independent and synergistic contributions of CGMB, HG-HCA, and FGRB in improving small-object detection, occlusion handling, and cross-scale robustness. This study demonstrates the potential of Hölder regularity and multi-domain fusion modeling in object detection, offering new insights for efficient visual modeling in complex indoor environments.

With the widespread application of high-speed and heavy-load railways, the real-time detection of track surface defects has become increasingly crucial. To address the challenges in rail defect detection, this study proposes the PerMSCA-YOLO model, which aims to overcome the limitations of traditional object detection models in multi-scale, small target, and complex background scenarios. By incorporating the lightweight FasterNet backbone network, a multi-scale convolutional attention module, and perceptual loss, the proposed model significantly enhances the detection accuracy and robustness of track defects. Experimental results show that PerMSCA-YOLO achieves an mAP@0.5 of 0.856, an F1-score of 0.79, and an inference frame rate of 142 FPS, demonstrating superior detection accuracy and real-time performance compared to other mainstream models like YOLOv8n. Furthermore, the model exhibits strong adaptability and efficiency when dealing with complex track defects, such as microcracks and corrosion patches, indicating its broad practical application potential. The innovative contribution of this research lies in its effective strategy for improving detection accuracy and real-time performance through multi-scale feature fusion and deep semantic alignment mechanisms, providing a solution that balances both precision and efficiency for defect detection in complex track environments, with substantial engineering application potential.

Traditional rehabilitation training for stroke patients with ankle joint issues typically relies on the expertise of physicians. However, when confronted with complex challenges, such as online decision-making or assessing rehabilitation progress, even seasoned experts may not anticipate all potential hurdles. A novel approach is necessary—one that effectively addresses these complexities without solely leaning on expert experience. Previous studies have introduced a rehabilitation assessment method based on fuzzy neural networks. This paper proposes a novel approach, which is a VR-aided ankle rehabilitation decision-making model based on a convolutional gated recurrent neural network. This model takes various inputs, including ankle dorsiflexion range of motion, angular velocity, jerk, and motion performance scores, gathered from wearable motion inertial sensors during virtual reality rehabilitation. To overcome the challenge of limited data, data augmentation techniques are employed. This allows for the simulation of five stages of rehabilitation based on the Brunnstrom staging scale, providing tailored control parameters for virtual training scenarios suited to patients at different stages of recovery. Experiments comparing the classification performance of convolutional neural networks and long short-term memory networks were conducted. The results were compelling: the optimized convolutional gated recurrent neural network outperformed both alternatives, boasting an average accuracy of 99.16% and a Macro-F1 score of 0.9786. Importantly, it demonstrated a strong correlation (correlation coefficient r > 0.9) with the assessments made by clinical rehabilitation experts, showing its effectiveness in real-world applications.

In the accelerated phase of urbanization, intelligent surveillance systems play an increasingly pivotal role in enhancing urban management efficiency, particularly in the realm of parking lot administration. The precise identification of small and overlapping targets within parking areas is of paramount importance for augmenting parking efficiency and ensuring the safety of vehicles and pedestrians. To address this challenge, this paper delves into and amalgamates cross-attention and multi-spectral channel attention mechanisms, innovatively designing the Criss-cross and Multi-spectral Channel Attention (CMCA) module and subsequently refining the CMCA-YOLO model, specifically optimized for parking lot surveillance scenarios. Through meticulous analysis of pixel-level contextual information and frequency characteristics, the CMCA-YOLO model achieves significant advancements in accuracy and speed for detecting small and overlapping targets, exhibiting exceptional performance in complex environments. Furthermore, the study validates the research on a proprietary dataset of parking lot scenes comprising 4502 images, where the CMCA-YOLO model achieves an mAP@0.5 score of 0.895, with a pedestrian detection accuracy that surpasses the baseline model by 5%. Comparative experiments and ablation studies with existing technologies thoroughly demonstrate the CMCA-YOLO model’s superiority and advantages in handling complex surveillance scenarios.

Machine vision is essential for intelligent industrial manufacturing driven by Industry 4.0, especially for surface defect detection of industrial products. However, this domain is facing sparse and imbalanced defect data and poor model generalization, affecting industrial efficiency and quality. We propose a perceptual capsule cycle generative adversarial network (PreCaCycleGAN) for industrial defect sample augmentation, generating realistic and diverse defect samples from defect-free real samples. PreCaCycleGAN enhances CycleGAN with a U-Net and DenseNet-based generator to improve defect feature propagation and reuse and adds a perceptual loss function and a capsule network to improve authenticity and semantic information of generated features, enabling richer and more realistic global and detailed features of defect samples. We experiment on ten datasets, splitting each dataset into training and testing sets to evaluate model generalization across datasets. We train three defect detection models (YOLOv5, SSD, and Faster-RCNN) with original data and augmented data from PreCaCycleGAN and other state-of-the-art methods, such as CycleGAN-TSS and Tree-CycleGAN, and validate them on different datasets. Results show that PreCaCycleGAN improves detection accuracy and rate and reduces the false detection rate of detection models compared to other methods on different datasets, demonstrating its robustness and generalization under various defect conditions.