To collect data of distributed sensors located at different areas in challenging scenarios through artificial way is obviously inefficient, due to the numerous labor and time. Unmanned Aerial Vehicle (UAV) emerges as a promising solution, which enables multi-UAV collect data automatically with the preassigned path. However, without a well-planned path, the required number and consumed energy of UAVs will increase dramatically. Thus, minimizing the required number and optimizing the path of UAVs, referred as multi-UAV path planning, are essential to achieve the efficient data collection. Therefore, some heuristic algorithms such as Genetic Algorithm (GA) and Ant Colony Algorithm (ACA) which works well for multi-UAV path planning have been proposed. Nevertheless, in challenging scenarios with high requirement for timeliness, the performance of convergence speed of above algorithms is imperfect, which will lead to an inefficient optimization process and delay the data collection. Deep learning (DL), once trained by enough datasets, has high solving speed without worries about convergence problems. Thus, in this paper, we propose an algorithm called Deep Learning Trained by Genetic Algorithm (DL-GA), which combines the advantages of DL and GA. GA will collect states and paths from various scenarios and then use them to train the deep neural network so that while facing the familiar scenarios, it can rapidly give the optimized path, which can satisfy high timeliness requirements. Numerous experiments demonstrate that the solving speed of DL-GA is much faster than GA almost without loss of optimization capacity and even can outperform GA under some specific conditions.

The proliferation of edge devices driven by advancements in Internet of Things (IoT) technology has intensified the challenge of achieving high-precision small target detection, as it demands extensive computational resources. This amplifies the conflict between the need for precise detection and the requirement for cost-efficiency across numerous edge devices. To solve this problem, this paper introduces an enhanced target detection algorithm, MSGD-YOLO, built upon YOLOv8. The Faster Implementation of CSP Bottleneck with 2 convolutions (C2f) module is enhanced through the integration of the Ghost module and dynamic convolution, resulting in a more lightweight architecture while enhancing feature generation. Additionally, Spatial Pyramid Pooling with Enhanced Local Attention Network (SPPELAN) replaces Spatial Pyramid Pooling Fast (SPPF) to expand the receptive field, optimizing multi-level feature aggregation for improved performance. Furthermore, a novel Multi-Scale Ghost Convolution (MSGConv) and Multi-Scale Generalized Feature Pyramid Network (MSGPFN) are introduced to enhance feature fusion and integrate multi-scale information. Finally, four optimized dynamic convolutional detection heads are employed to capture target features more accurately and improve small target detection precision. Evaluation on the VisDrone2019 dataset shows that compared with YOLOv8-n, MSGD-YOLO improves mAP@50 and mAP@50–95 by 14.1% and 11.2%, respectively. In addition, the model not only achieves a 16.1% reduction in parameters but also attains a processing speed of 24.6 Frames Per Second (FPS) on embedded devices, thereby fulfilling real-time detection requirements.

When forgery techniques can generate highly realistic videos, traditional convolutional neural network (CNN)-based detection models often struggle to capture subtle forgery features and temporal dependencies. Most existing models focus on feature extraction from static frames, neglecting the temporal correlation in videos, which decreases accuracy in detecting dynamic forged videos. Furthermore, the ability to detect localized forgery features remains insufficient. To address these limitations, we propose a deep forgery detection framework named BiLSTM Multi-Head Self-Attention Network (BMNet). By leveraging the Bi-Directional Long Short-Term Memory Network (BiLSTM) for modeling temporal dependencies between video frames, and the Multi-Head Self-Attention Mechanism (MHSA) for capturing features from different regions of an image, BMNet more effectively identifies dynamic and local forgery features. In our experiments, we extract features from 68 facial landmarks of each video frame and evaluate detection performance on four datasets: UADFV, FF++, Celeb-DF and DFDC. The results show significant improvements over traditional methods. We further validate the necessity of each network component through ablation studies, demonstrating that BMNet achieves accuracies of 95.54%, 92.18%, 80.20% and 84.72% on the FF++, UADFV, Celeb-DF and DFDC datasets, respectively, indicating its superior performance in deep forgery detection.