Seal identification method and system fusing data enhancement and deep learning

• Main Authors: CHEN SHUAIPIAO, WANG XIAOGUANG, JIANG TINGTING, WANG WENHUI, ZHANG ZHIJIAN, MA FEICHENG, HUO CHAOGUANG
• Format: Patent
• Language: Chinese; English
• Published: 02.08.2024
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; PHYSICS

The invention discloses a seal recognition method fusing data enhancement and deep learning, and the method comprises the steps: obtaining a seal image, and carrying out the recognition and marking of the seal image; performing data enhancement on the marked seal image to obtain enhanced image data, and dividing the enhanced image data into a training set, a verification set and a test set; constructing a ViT deep learning model, the ViT deep learning model being used for identifying the seal image, training the ViT deep learning model by using the training set, and performing verification by using the verification set to obtain a trained ViT deep learning model, and finally, evaluating the performance of the trained ViT deep learning model by using the test set, and selecting the ViT deep learning model of which the performance meets the requirement. And adopting the obtained ViT deep learning model to identify a to-be-identified seal image. According to the method, the problem that the neural network model