Research on image classification and recognition technology based on depth learning

A new image classification algorithm based on multi-feature extraction and improved SVM classifier is analyzed theoretically and numerically. Image object classification and detection are two basic problems in computer vision, the key is how to fully mine a large amount of redundant information, and on this basis to make effective feature supplement, thus, the classification accuracy of the image is reduced. To solve this problem, this paper proposes an improved SVM algorithm by combining image classification and synthesis technology. This paper proposes a video sequence segmentation method based on PCA, which can describe the content characteristics of the image comprehensively, and use PCA to extract conversion features and eliminate redundancy, the experimental results show that the algorithm is effective and feasible.