Feature Extraction and Recognition Based on Integrated Material Painting Images

Abstract

With the rise of digital art, the demand for classification and recognition technology of images is increasing. The purpose of this study is to improve the accuracy of feature extraction and classification recognition in Chinese painting images. A method combining multicolor gamut texture analysis and block color feature extraction is introduced. The multiresolution grayscale co-occurrence matrix technology is applied to enhance the expression ability of feature vectors. Form the results, the average accuracy and recall were improved by 12.2 and 14% respectively compared to traditional grayscale co-occurrence matrix methods. In terms of noise resistance, the algorithm proposed in the study showed a 7 and 7.5% decrease in average accuracy and recall under 30dB noise conditions, which was significantly better than traditional methods, proving the significant advantage of the algorithm in terms of robustness. In summary, the feature extraction method proposed in the research has effectively improved the accuracy and robustness of Chinese painting image classification. This provides a new technological path for image analysis in the field of digital art, laying the foundation for the development of art image processing technology.