Intelligent Classification Method of Basic Elements of Chinese Ink Painting Based on a Convolution Neural Network

Facing the problem that current experience-based expert, physical and chemical analysis and identification methods lack in-depth learning of samples, resulting in inaccurate classification results, an intelligent classification method of basic elements of Chinese ink painting based on a convolutional neural network is proposed. The convolutional neural network structure is designed, and the basic elements of the seal, house, moss point, Figure and branch are analyzed. Combined with the deep learning algorithm, the activation function is initialized to avoid the saturation effect and improve the convergence speed. According to the clustering principle based on the deep learning algorithm, samples with membership less than 1 in clustering are selected to avoid the adverse impact of disordered elements on classification. Combined with the loss function curve, the classification results are output. Taking the world art database as an example, this paper makes an example analysis. According to the analysis results, the highest classification accuracy of this method is 0.9856, which has accurate classification results.