Brain-computer Collaborative Method for Low Quality Video Object Detection

Target detection technology is crucial in various domains. However, traditional detection methods, including manual identification and computer vision (CV), face limitations when dealing with low-quality videos. Brain-computer interface (BCI) technology provide a promising alternative, but still encounter challenges such as high false alarm rates and susceptibility to environmental interference. This paper proposes a brain-computer fusion approach that integrates CV with BCI through a novel video stimulation paradigm. Additionally, an adaptive collaborative framework is introduced, employing a dual-stream fusion architecture with a dynamic adjustment mechanism guided by CV. This method combines the strengths of CV in image processing and target localization with the cognitive advantages of BCI to improve detection performance. The experimental outcomes demonstrate that the proposed fusion approach achieved higher F1-scores and accuracy by substantially reducing false alarm rates. This work not only introduces novel technical insights into the field of video target detection but also highlight the potential of human-machine collaboration to address challenges in automated detection.