In order to solve the problems of time-consuming and low classification accuracy of feature extraction in traditional painting feature classification methods, a new method based on stochastic forest algorithm is proposed. The colour feature of painting is converted into HSV component, and the original LBP value of the painting image is converted into 59-dimension feature vector by uniform mode to extract the painting texture feature. The wavelet transform method is used to obtain the high and low frequency band signal of painting features, and the noise reduction of painting features is comple...