In order to improve the accuracy of image recognition, a new edge detection image recognition algorithm is proposed. This algorithm first uses the Canny operator to identify the edge pixels of the image, and then calculates the gradient of each valid pixel. The normalized histogram is established by the obtained pixel gradient sequence. In the recognition process of the deep learning-based handwritten calligraphy font recognition algorithm, the image processing methods such as projection method are used to locate and segment the Chinese characters in the calligraphy work image, and then use the GoogLeNet Inception- v3 model and ResNet-50 residual network for book style recognition and glyph recognition. Artificial intelligence technology can generate fast and rich teaching auxiliary information for different students or different creative intentions, and intuitively convey teaching goals in art teaching. Improve students' learning efficiency.