ARF-YOLOv8: a novel real-time object detection model for UAV-captured images detection

Journal of Real-Time Image Processing, Год журнала: 2024, Номер 21(4)

Опубликована: Июнь 4, 2024

Язык: Английский

---

FusionU10: enhancing pedestrian detection in low-light complex tourist scenes through multimodal fusion

Frontiers in Neurorobotics, Год журнала: 2025, Номер 18

Опубликована: Янв. 10, 2025

With the rapid development of tourism, concentration visitor flows poses significant challenges for public safety management, especially in low-light and highly occluded environments, where existing pedestrian detection technologies often struggle to achieve satisfactory accuracy. Although infrared images perform well under conditions, they lack color detail, making them susceptible background noise interference, particularly complex outdoor environments similarity between heat sources features further reduces To address these issues, this paper proposes FusionU10 model, which combines information from both visible light images. The model first incorporates an Attention Gate mechanism (AGUNet) into improved UNet architecture focus on key generate pseudo-color images, followed by using YOLOv10. During prediction phase, optimizes loss function with Complete Intersection over Union (CIoU), objectness (obj loss), classification (cls thereby enhancing performance network improving quality feature extraction capabilities through a feedback mechanism. Experimental results demonstrate that significantly improves accuracy robustness scenes FLIR, M3FD, LLVIP datasets, showing great potential application challenging environments.

Язык: Английский

---

Infrared maritime small target detection network based on attention and partial learning convolution

Infrared Physics & Technology, Год журнала: 2025, Номер unknown, С. 105748 - 105748

Опубликована: Фев. 1, 2025

Язык: Английский

---

SAM-YOLO: An Improved Small Object Detection Model for Vehicle Detection

Опубликована: Фев. 27, 2025

Vehicle detection using computer vision plays a crucial role in accurately recognizing and responding to various road conditions, targets, signals, particularly within autonomous driving technology. However, traditional vehicle algorithms suffer from slow speed, low accuracy, poor robustness. To address these challenges, this paper proposes the simple attention mechanism-you only look once (SAM-YOLO) algorithm. SAM-YOLO incorporates mechanism into YOLOv7 network, allowing for capture of more detailed information without introducing additional parameters. In study, we experimentally redesigned backbone network by replacing redundant part layer with C3 module, resulting improved model performance while maintaining accuracy. The experimental results show that algorithm performs excellently several evaluation metrics under conventional especially outperforming other accuracy mean average precision values. tests on ExLight dataset facing extreme lighting similarly demonstrated optimal capabilities, terms robustness when dealing complex variations. These findings emphasize potential real-time accurate target tasks, environments highly variable conditions.

Язык: Английский

---

A multiscale and cross-level feature fusion method for remote sensing image target detection

Advances in Space Research, Год журнала: 2025, Номер unknown

Опубликована: Март 1, 2025

Язык: Английский

---