Abstract.To address the issue of low tracking accuracy for occluded targets in complex scenes, we propose a multi-object tracking algorithm (GradsSORT) based on pseudo-gradient. First, targets are categorized into different gradient layers according to their degree of occlusion. Then, a hierarchical association strategy is adopted during data matching, prioritizing the association of detection boxes in lower gradient layers. This approach helps remove occluders for occluded targets, making it more effective for tracking tasks in crowded occlusion scenarios. To verify the effectiveness of the proposed … Read more →

Abstract.In view of the problem that the workshop of intelligent terminal structural parts is faced with the disturbance caused by urgent orders affecting the production efficiency and delivery punctuality of the workshop, combined with the initial scheme minimization index and the order insertion rescheduling delay minimization index, a mathematical model of order insertion scheduling in the production workshop of intelligent terminal structural parts is established; And genetic algorithm is used to solve it, dynamically adjusting the probability of non-uniform crossover and non-uniform mutation to … Read more →

Abstract.Objectives: A schizophrenia patient model based on Persistent homology and multilayer brain networks combines full-frequency band information to mine and categorize key brain regions, providing a decision-making basis for patient treatment. Methods: Using the information of 198 subjects from Huilongguan Hospital in Beijing who met the requirements of this study, the network features based on the constructed multilayer brain network were mined by Persistent homophony and traditional methods of mining multi-frequency band information of subjects, and finally analyzed by using a deep learning … Read more →

Abstract.Federated learning for internet of vehicles builds an intelligent transportation system with real-time responsiveness and intelligent collaborative training of high-quality models by integrating traffic data between vehicle nodes, roadside units, and infrastructure. As the internet of vehicles architecture continues to expand, frequent gradient data interactions between roadside units and vehicle nodes lead to increased uplink channel load and communication delay in federated learning systems. To alleviate the communication delay problem, existing works propose gradient quantization algorithms … Read more →

Abstract.In Simultaneous Localization and Mapping (SLAM) algorithms, the robustness of single sensors is relatively poor. To address this issue, this study develops a multi-sensor fusion platform that integrates data from LiDAR, stereo cameras, and IMUs, and employs advanced algorithms such as ORB-SLAM2 and Cartographer for efficient data processing and map construction. For point cloud data, we apply the Iterative Closest Point (ICP) algorithm for point-to-point optimization, effectively solving the issue of large map accumulation errors in traditional LiDAR-based SLAM algorithms, which rely solely on … Read more →

Abstract.The ring signature ensures the anonymity of the signer and the unforgeability of the message during communication. However, the presence of sensitive data in the message may expose the signer's identity to some extent. To address this issue, we propose an efficient lattice-based ring signature scheme with range proofs. The scheme uses a commitment mechanism to ensure the signer conceals sensitive data while allowing the verifier to effectively validate the data's range. Additionally, a tracking algorithm is introduced to mitigate abusive behavior by revealing the identity of the abuser. … Read more →

Abstract.With the rapid development of electronic manufacturing industry, printed Circuit Board (PCB) is the core hub of electronic products, and its quality inspection plays a decisive role in ensuring the performance and reliability of products. In order to overcome the shortcomings of weak interactivity and low visualization of traditional PCB board defect detection methods, this study designed a highly interactive and intuitive PCB board defect detection system based on YOLOv5 with the help of PyQt toolkit. The system has a UI with complete functions and stable performance, which includes three … Read more →

Abstract.Open-World Object Detection (OWOD) aims to address the core challenges that traditional detection models cannot dynamically adapt to unknown objects and continuously learn new categories. Most existing methods rely on visual similarity to discover new objects, yet they still face issues such as semantic disconnection and catastrophic forgetting in complex cross-domain scenarios, such as medical imaging and autonomous robotics. This paper proposes a causality-driven open-world object detection framework. By decoupling the functional causal features and the appearance features of objects, it … Read more →

Abstract.Large language models (LLMs) have achieved remarkable success in natural language processing (NLP) tasks. However, fine-tuning LLMs using private datasets raises significant privacy concerns, as models can inadvertently memorize sensitive information. Differentially Private Stochastic Gradient Descent (DP-SGD) provides a mathematically rigorous solution but suffers from high computational overhead, slow convergence, and excessive privacy budget consumption, making it impractical for large-scale models. To address these challenges, we propose an efficient differentially private fine-tuning … Read more →

Abstract.With the rapid development of artificial intelligence technology, Transformer architecture has become the core framework of natural language processing (NLP) and multimodal domain. In this paper, the fusion strategies of Large Language Model (LLM) and Visual Language Model (VLM) with Transformer architecture are deeply studied. This paper first introduces the basic principles and characteristics of Transformer architecture, LLM and VLM models, and then makes a comprehensive analysis of the advantages and challenges of different fusion strategies, and demonstrates the practical application … Read more →

Abstract.In recent years, the advancement of cableless seismic data acquisition devices, particularly nodal seismographs, has revolutionized seismic exploration. However, traditional "blind acquisition" methods face significant limitations, including the inability to enable real-time monitoring, challenges in ensuring data transmission stability, and the inefficiency of device retrieval in large-scale exploration projects. To address these issues, this paper proposes an Efficient Com-pressed Sensing-Driven Wireless Multi-Hop Seismic Data Transmission Network (WMDTN), which inte-grates a hierarchical … Read more →