Watercolor Nationality Recognition in the Context of Concrete Painting Based on Texture Recognition and Heterogeneous Database

This paper is based on texture recognition and heterogeneous database of concrete painting in the context of watercolor nationality recognition, image-based classification and recognition, and the use of gray-level co-occurrence matrix-based texture feature extraction. Taking the 8-bit grayscale image as the research object, the image is scanned in 4 directions (o, 450, 90, 1350) in sequence, and the aesthetic concept of concrete expression painting is considered. Analyze the experience and works of the concrete expression masters based on the heterogeneous database and establish the friendship cooccurrence matrix respectively to extract the texture feature parameters, and then establish the BP neural network model, and use the extracted texture feature values to train the BP neural network model to make the network model Able to identify the content of the image.