About this document
Fastc 2 F by 22119209 is a document available to read on EtoBox.
The paper presents FDM-YOLO, a lightweight model based on YOLOv8, designed to enhance small target detection in drone imagery and traffic scenarios while maintaining real-time performance. Key innovations include a modified detection head, the Fast-C2f structure, dynamic upsampling, and an EMA attention mechanism, leading to a 38% reduction in parameters and improved accuracy from 38.4% to 42.5% on the Visdrone dataset. This approach effectively balances accuracy and efficiency, making it suitable for deplo
- Author
- 22119209
- Language
- EN