Research on Digital Training System of University Sports Based on SIFT Algorithm

In order to enhance the humanization and diversification of human-computer interaction, a digital training system for school sports was designed. Due to the large amount of data and high computational complexity of the content image classification retrieval method, an image classification method based on the SIFT algorithm is proposed and applied to sports image classification. In addition, we can also use the SIFT algorithm to find the image corresponding to the image to be tested, give full play to the advantages of the SIFT algorithm recognition technology, and use the output information to complete the control work of the practical application. After the method extracts the feature points from the image, it uses the DBS which can algorithm and the K-Mean algorithm to analyze the feature data, so as to obtain the data that best reflects the image characteristics, and then use these data to classify the image.