Analysis method for generating characters based on expression recognition and psychological tests

The invention relates to an analysis method for generating characters based on expression recognition and psychological tests, and belongs to the field of convolutional neural networks. The method comprises the steps of capturing face image information, processing the captured face image; extracting features of a face image through two convolutional neural networks, performing multi-classificationon the obtained facial expression features through a support vector machine to obtain expression psychological features, performing dual analysis through the expression psychological features and psychological test features, and dividing the character analysis into three types: a rational intelligence type, an emotion type and an intention type. According to the method, the facial expression is identified through the convolutional neural network, the risk of over-fitting can be better reduced, and more accurate expression features are obtained. Characteristics of a psychology test are added in the personality analysi