Method, system, device, medium and program product for processing handwritten signatures

Abstract

The present disclosure provides a handwritten signature processing method, system, device, medium and program product, the processing method comprising: obtaining a handwritten signature image to be processed; processing the handwritten signature image through a trained signature processing model to obtain a clarity discrimination result and a general feature vector of the handwritten signature image, wherein the signature processing model comprises an image branch, a text branch, a clarity classification head and a gated neural network; selecting a post-processing path according to the clarity discrimination result to obtain a signature text. The present disclosure processes the handwritten signature image through the signature processing model to obtain the clarity discrimination result of the handwritten signature image, and then selects the post-processing path according to the clarity discrimination result to obtain the signature text. The clarity discrimination of the handwritten signature image is set in the signature processing model, which improves the detection efficiency, reduces the complexity of the recognition method, improves the recognition efficiency, can adapt to more scenes, and reduces resource consumption.

Images