A Research on Image Recognition and Classification Based on Traditional Machine Learning and Deep Learning

With the rapid development of smart devices, a huge amount of images are appearing in people's lives, so the processing of images, especially recognition and classification processing, becomes more and more important. This paper aims to provide researchers with an overview of current traditional machine learning (ML) and deep learning (DL) in solving image recognition and classification problems, specifically analyzing architectures such as KNN, SVM, Random forest algorithms, and CNN Architecture, and finding available data training results for comparative analysis. In summary, it is found that deep learning algorithms perform well in handling large-scale complex datasets and automatic feature extraction, and are applied to most image recognition and classification problems, but the traditional machines have obvious advantages of accessibility and speed in handling small-scale data, and deep learning algorithms are still not an effective choice in some special cases or resource-limited environments. At the end of the article, the full text is summarized and outlook.