一种基于脑-机信号融合的目标检测方法
By constructing a brain-computer signal fusion model that integrates multi-head attention mechanism and cross-modal knowledge distillation, and combining EEG and image features, the problem of insufficient accuracy and robustness of target detection in UAV aerial images is solved, achieving high-precision and robust target detection results.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- NAT UNIV OF DEFENSE TECH
- Filing Date
- 2023-05-08
- Publication Date
- 2026-07-17
AI Technical Summary
Target detection accuracy and robustness in drone aerial images are low; computer vision methods struggle to effectively identify targets in complex backgrounds; EEG single-modal detection is susceptible to environmental noise interference; and existing combined methods are under-researched in this field.
A brain-computer signal fusion model based on multi-head attention mechanism and cross-modal knowledge distillation is constructed. Target detection is performed by fusing EEG features and image features. The model utilizes MCGRAM to process the frequency-space-time features of EEG data and EfficientNet to process image data. Combined with a multimodal feature fusion module, it achieves high generalization, strong robustness, and high accuracy in target detection.
High-precision and robust target detection was achieved in drone aerial images, improving the automatic target detection performance in the monitoring system and effectively complementing the advantages and disadvantages of computer vision and EEG.
Smart Images
Figure CN116524380B_ABST