Mathematical Modeling Optimization of Recognition Algorithm Based on Multi Feature Extraction

The research on recognition algorithm of multi feature extraction is to find the best method to extract features from images. Researchers used different algorithms and experiments to find the best algorithm. The algorithm adopts the idea of classification and discrimination in the matching and discrimination stage. While inputting the Chinese character image, the input structure characteristics of the Chinese character image are recorded. On this basis, the stroke characteristics of the input unknown Chinese characters are obtained. From the stroke characteristics, the unknown Chinese characters can be roughly classified, which can narrow the range of Chinese characters in the discrimination. Some of these algorithms include: SIFT (scale invariant feature transformation) - this algorithm is used in many applications, such as face detection, image retrieval and target tracking. It can be applied to any type of image, including video sequences. Sift uses a set of parameters based on scale invariance, so it can process any type of image or video sequence without distortion.