A Text Detection and Recognition Method Based on PSENet and CRNN

Text detection refers to detecting, for a given image, the coordinates of the text regions within it, and text recognition refers to identifying specific characters from the detected text boxes and optimizing the recognition results. In this paper, a text detection and recognition method based on PSENet and CRNN is designed for book name recognition in conjunction with the application scenario of automatic book return in self-service libraries. The PSENet text detection algorithm is used to obtain the text box of the book cover, and the CRNN+CTC text recognition algorithm is used to perform text recognition on the name of the book cover, achieving a better recognition effect.