Target lightweight detection method and system based on attention feature enhancement
By designing a lightweight backbone network, a dedicated attention mechanism, and an adaptive feature fusion module, combined with hybrid precision quantization and channel pruning, the problem of high computational complexity in existing models is solved, enabling real-time visible light target detection on embedded devices.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2025-09-19
- Publication Date
- 2026-06-05
AI Technical Summary
Existing deep learning object detection models are too computationally complex and have too many parameters, making it difficult to achieve real-time detection on resource-constrained embedded cameras and mobile terminals, especially in terms of insufficient detection accuracy under complex and variable visible light environments.
We design a lightweight detection method based on attention feature enhancement, including a lightweight backbone network, a dedicated attention mechanism, an adaptive feature fusion module, and a detection head. We combine mixed precision quantization, knowledge distillation, and channel pruning to compress the model and reduce computational resource requirements.
While ensuring detection accuracy, it significantly reduces the computational resource requirements, enabling real-time detection of targets in visible light images, and is suitable for embedded cameras and mobile terminals.
Smart Images

Figure CN121147496B_ABST
Abstract
Citation Information
Patent Citations
Lightweight target detection method and device
CN118762162A