New abnormal target detection method based on brain-computer fusion cognition and decision

Abstract

The invention discloses a novel abnormal target detection method based on brain-computer fusion cognition and decision. The method comprises the following steps: 1, calculating a classification value of an abnormal target image through a computer vision target detection algorithm; 2, designing a fast sequence visual presentation experiment to induce an electroencephalogram signal corresponding to an abnormal target image, performing feature extraction and classification on the electroencephalogram signal, and calculating a classification value of the abnormal target image; 3, evaluating the classification performance of the computer and the human brain on the abnormal target image, and calculating a trust weight; and 4, establishing a D-S evidence theory brain-computer fusion cognition and decision-making model, and according to classification values of a trust weight fusion computer and a human brain, calculating a brain-computer fusion classification value of whether an abnormal target exists in the image, so as to obtain an abnormal target detection result. The method can fully fuse the decision information of the computer and the human brain, reduces the decision contradiction between the computer and the human brain, improves the brain-computer fusion performance, and effectively solves the problem of low accuracy of abnormal target detection.

Description

technical field [0001] The invention belongs to the interdisciplinary research field of brain-computer interface, computer vision and intelligent information fusion. Background technique [0002] Target detection is to find out the target feature expression in the image to distinguish the target from the non-target. The development of machine learning and deep learning has promoted the development of computer vision object detection, and the accuracy has been greatly improved. However, for abnormal targets in low-visibility scenes such as night and snowy days, the target imaging quality is poor, and the computer lacks sufficient recognition ability to meet the accuracy requirements. Human beings have strong recognition ability, can obtain key visual information in an image at a glance, and can quickly detect objects of interest in an image. Brain-computer interface is a communication and control technology that allows people to communicate directly with the outside world t...

AI Technical Summary

Problems solved by technology

[0003] In recent years, researchers at home and abroad have done more research on abnormal target detection. Their research mainly focuses on the improvement of computer abnormal target detection models, feature extraction and classification of EEG signals in the target state, computer and human brain. These studies have made good progress, but there are still many problems: (1) The accuracy of computer detection is very low

The image features of the abnormal target itself change greatly, which increases the difficulty of image feature analysis and reduces the accuracy of computer detection

(2) There are decision-making contradictions in the fusion cognition of human brain and computer

The reasons for the false detection of the computer and the human brain are different, and the detection results of the computer and the human brain are inconsistent, which leads to the contradiction between the decision-making of the computer and the human brain

Images