Combining the advantages of conventional and event cameras for robust visual tracing has drawn extensive interest. However, existing tracking approaches heavily engage in complex cross-modal fusion modules, leading to higher computational complexity and training challenges. Besides, these methods generally ignore the effective integration of historical information, which is crucial to grasping the change in the target’s appearance and motion trends. Given the recent advancements in Mamba’s long-range modeling and linear complexity, we explore its potential in addressing the above issues in RGBE tracking tasks. Specifically, we first propose an efficient fusion module based on Mamba, which utilizes a simple gate-based interaction scheme to achieve effective modality-selective fusion. This module can be seamlessly integrated into the encoding layer of prevalent Transformer-Based backbones. Moreover, we further present a novel historical decoder that leverages Mamba’s advanced long sequence modeling to effectively capture the target appearance changes with autoregressive queries. Extensive experiments show that our proposed approach achieves state-of-the-art performance on multiple challenging short-term and long-term RGBE benchmarks. Besides, the effectiveness of each key Mamba-Based component of our approach is evidenced by our thorough ablation study.Code will be released at: https://github.com/scy0712/MamTrack

Most existing researches achieve image matting via predefined trimap or saliency-guided intermediate segmentation variants, and their essence is providing localization hints for the matting targets. In this paper, we propose our Localization Hints Exploration object matting model (LHEMatt) to estimate alpha mattes based on simple and flexible localization hints. We design our location constraints enhancement module (LCE) to exploit more efficient and effective hints for alpha matte prediction. A well-designed foreground details extension module (FDE) is also presented to integrate target locations with rich boundary details. Besides, to accommodate more computer vision tasks and practical applications, we extend the matting scope to the multi-object foreground and explore separated localization hints to disassemble several targets for object-level matting. Based on this motivation, we construct a new dataset consisting of 757 different delicate alpha mattes (Mobjects-757), and most of them contain two or more objects. As the first object matting inspiration, we perform extensive experiments to demonstrate the proposed localization hints exploration model and evaluate different approaches on the newly-created Mobjects-757 matting dataset.

Underwater object detection (UOD) is crucial for monitoring marine ecosystems, underwater robotics, environmental protection, and autonomous underwater vehicles (AUVs). Despite progress, many models struggle under real-world conditions due to poor visibility, dynamic lighting, and domain shifts. Traditional methods like Faster R-CNN are computationally expensive, while YOLO-based models suffer in challenging underwater scenarios. The scarcity of large-scale annotated datasets further limits model generalization. To address these challenges, we introduce UOD-SZTU-2025, a new dataset of 3,133 high-quality underwater images, sourced primarily from video platforms. The dataset is used in EFCWM (Enhanced Feature Correction and Weighting Module) to extract and refine a feature material library for detection targets. We propose EFCWM-Mamba-YOLO, a lightweight, real-time detection model designed to enhance feature representation and adapt to diverse underwater environments. The EFCWM module incorporates domain adaptation for improved robustness. Additionally, a two-stage training strategy first trains on a source domain and fine-tunes with limited target domain samples to enhance generalization. Experiments show our approach surpasses existing lightweight UOD models in accuracy, real-time performance, and robustness. Our dataset, model, and benchmark establish a strong foundation for future UOD research. The dataset for EFCWM-Mamba-YOLO is available at https://github.com/wojiaosun/UOD-SZTU-2025.

Unmanned aerial vehicles could accurately accomplish complex navigation and obstacle avoidance tasks under external control. However, enabling unmanned aerial vehicles (UAVs) to rely solely on onboard computation and sensing for real-time navigation and dynamic obstacle avoidance remains a significant challenge due to stringent latency and energy constraints. Inspired by the efficiency of biological systems, we propose a fully neuromorphic framework achieving end-to-end obstacle avoidance during navigation with an overall latency of just 2.3 milliseconds. Specifically, our bio-inspired approach enables accurate moving object detection and avoidance without requiring target recognition or trajectory computation. Additionally, we introduce the first monocular event-based pose correction dataset with over 50,000 paired and labeled event streams. We validate our system on an autonomous quadrotor using only onboard resources, demonstrating reliable navigation and avoidance of diverse obstacles moving at speeds up to 10 m/s.

Smart contracts on the open network (TON) have become vital in Internet of Things (IoT) applications due to their low latency and high scalability. However, the unique architectural features of TON introduce specialized vulnerabilities that existing tools fail to address comprehensively. In this letter, we propose a novel defect detection framework that combines large language models (LLMs) for automated defect discovery with a locatable call graph for precise and efficient code analysis. Our method identifies four new types of TON-specific defects: 1) Ignore Errors Mode Usage; 2) Premature Acceptance; 3) Pseudo Deletion; and 4) Improper Jetton Refund. Evaluated on 1640 real-world smart contracts written in FunC and Tact, the framework uncovers 669 defects, with an average of one defect every 2.45 code segments. The detection achieves an average F1 score of 99.75% for FunC and 100% for Tact contracts. Additionally, our approach demonstrates lightweight computational overhead, consuming only 12.6 MB of memory and achieving a mean response time of 0.05 s. These results highlight the accuracy, efficiency, and practicality of our framework for securing TON-based smart contracts in IoT ecosystems.

In this paper, we focus on diversifying the Video Moment Retrieval (VMR) model into more scenes. Most existing video moment retrieval methods focus on aligning video moments and queries by capturing the cross-modal relationship, which largely ignores the cross-instance relationship behind the representation learning. Thus, they may easily get trouble into the inaccurate cross-instance contrastive relationship in the training process: 1) Existing approaches can hardly identify similar semantic content across different scenes. They incorrectly treat such instances as negative samples (termed faulty negatives), which forces the model to learn the features from query-irrelevant scenes. 2) Existing methods perform unsatisfactorily in locating the queries with subtle differences. They neglect to mine the hard negative samples that belong to similar scenes but have different semantic content. In this paper, we propose a novel robust video moment retrieval method that prevents the model from overfitting the query-irrelevant scene features by accurately capturing both the cross-modal and cross-instance relationships. Specifically, we first develop a scene-independent cross-modal reasoning module that filters out the redundant scene contents and infers the video semantics under the guidance of query information. Then, the faulty and hard negative samples are mined from the negative ones and calibrated for their contribution to the overall loss in contrastive learning. We validate our contributions through extensive experiments on cross-scene video moment retrieval settings, where the training and test data are from different scenes. Experimental results show that the proposed robust video moment retrieval model can effectively retrieve target videos by capturing the real cross-modal and cross-instance relationships.

Recently, as one of the most popular covert communication technologies, generative steganography has received ever-increasing attention due to its promising performance against sophisticated steganalysis tools. However, it is quite difficult for the existing generative steganographic approaches to find a good tradeoff between hiding capacity and extraction accuracy, mainly due to the small capacity of their hiding spaces. To overcome this shortcoming, a Progressive Generative Steganography (PGS) network architecture is proposed to hide a secret message during the progressive image generation process to realize secure covert communication. Specifically, we first propose a robust Secret-to-Noise (S2N) mapping method to encode the secret message as a set of noise maps. Then, guided by these noise maps, a set of corresponding images ranging from low resolution to high resolution are progressively generated by the Single Generative Adversarial Networks (SINGAN). Consequently, a large-sized secret message can be hidden in the finally generated high-resolution image, since a set of high-capacity hiding spaces can be provided by the process of progressive image generation. Moreover, to improve the quality of image generation and the accuracy of secret message extraction, a Dense Secret-Feature Connection (DSFC) strategy is designed and integrated into the proposed PGS network architecture. Extensive experiments demonstrate that the proposed PGS outperforms the existing approaches in the aspects of both hiding capacity and message extraction, while maintaining promising anti-detectability and imperceptibility for covert communication.

UAV gesture recognition, a novel human-computer interaction form, offers an intuitive approach to controlling UAVs in various environments. However, there is a lack of comprehensive datasets for AI-powered UAV gesture recognition. This paper contributes in several ways: (i) We introduce MD-UHGRD, a unique UAV static gesture dataset with 20, 000 images and annotations, collected from a diverse group of participants in different environmental conditions. This dataset is expected to bridge a significant gap in UAV gesture recognition algorithms. (ii) We propose SA-YOLO, a multifunctional UAV gesture recognition method that not only enables gesture recognition but also includes face and pedestrian tracking, optimizing UAV control in complex scenarios. SA-YOLO incorporates the Spatial Asymptotic Feature Pyramid Network (SAFPN), Scale Pyramid Pooling with Cross Stage Partial Networks Convolution (SPPCSPC), and Space-to-Depth Convolution (SPD-Conv). (iii) Extensive evaluation of SAYOLO on MD-UHGRD establishes it as a benchmark in this domain. Our method demonstrates high accuracy, processing speed, and a compact model size, achieving a 93.2% mean Average Precision (mAP) with 10.3 million parameters and 48 frames per second (FPS). Among competing models, SA-YOLO not only achieves the highest mAP but also maintains a balance in model size and FPS. The database and code are available at: https://github.com/ijcnn2024/SA-YOLO.

Ensuring fetal health during pregnancy is critically dependent on precise Fetal Heart Rate (FHR) monitoring. A major challenge in this area is the limited availability of labeled FHR data, which poses a barrier to developing reliable automated analysis systems. To address this gap, our study introduces CTGGAN, a novel method employing Generative Adversarial Networks (GANs) to create synthetic, high-quality FHR signals. Specifically, CTGGAN integrates self-attention and residual modules within a Conditional GAN framework, fine-tuned to replicate the complex patterns characteristic of FHR data accurately. A notable feature of CTGGAN is its effective loss function, which combines Wasserstein distance with a gradient penalty to ensure training stability and enhance the authenticity of the generated signals. In performance metrics, Our method demonstrates the highest signal fidelity and distribution similarity, across five key measures: 0.215 maximum mean deviation (MMD), 0.012 sliced Wasserstein distance (SWD), 4.821 percent root mean square difference (PRD), 5.621 relative entropy (RE), and 0.614 Frechet distance (FD). This advancement in generating realistic FHR data with CTGGAN addresses critical issues like data insufficiency and class imbalance, thus advancing the field of prenatal healthcare technology. The code for CTGGAN is available at https://github.com/ijcnn2024/CTGGAN.

Cardiovascular diseases (CVDs) are the leading cause of global morbidity and mortality, necessitating the precise and continuous monitoring of blood pressure for proactive management. Our study presents the AMRUNet: a novel network designed exclusively for PPG-only, noninvasive, cuff-less blood pressure estimation. The network innovates upon the U-Net architecture, integrating a MultiRes Block for detailed multi-scale feature fusion and a residual block to mitigate the issue of vanishing gradients. An attention mechanism is further employed to selectively enhance salient features within the PPG signal. Our PPG-only AMRUNet demonstrates exceptional performance in translating PPG data into accurate ABP waveforms, achieving mean absolute errors (MAE) that comply with the standards of both the British Hypertension Society (BHS) and the Association for the Advancement of Medical Instrumentation (AAMI). Our method demonstrates highest MAE for both systolic blood pressure (SBP) and diastolic blood pressure (DBP), achieving a 2.85 MAE for SBP and 1.79 MAE for SBP among competing models. The model’s proficiency in precisely estimating systolic and diastolic blood pressure, along with its ability to reconstruct continuous ABP waveforms, contributes to reliable and trustworthy medical decision-making systems. The code for AMRUNet can be accessible at https://github.com/ijcnn2024/AMRUNet.

The advancement of machine vision systems necessitates efficient and accurate signal reconstruction methods to enhance real-time perception and decision-making capabilities. This paper introduces a Generalized Backtracking Regularization Adaptive Matching Pursuit (GBRAMP) algorithm, designed to reconstruct signals within machine vision systems using compressed sensing techniques. The GBRAMP algorithm improves upon existing methods by incorporating regularization for enhanced atom selection and a backtracking approach to accurately estimate sparsity, addressing the limitations of traditional convex optimization, greedy, and Bayesian reconstruction algorithms. The paper provides a comparative analysis of the GBRAMP algorithm against other prominent reconstruction techniques. Experimental results validate the GBRAMP algorithm's improved performance in terms of both reconstruction accuracy and computational speed, making it a competitive solution for the next generation of machine vision systems.