Training method and training device for amplifying training data of document analysis model

• Main Authors: CHEN CHANGSHENG, ZHANG SHUZHENG, ZHU GANG
• Format: Patent
• Language: Chinese; English
• Published: 25.03.2022
• Subjects: CALCULATING; COMPUTING; COUNTING; HANDLING RECORD CARRIERS; IMAGE DATA PROCESSING OR GENERATION, IN GENERAL; PHYSICS; PRESENTATION OF DATA; RECOGNITION OF DATA; RECORD CARRIERS

The invention discloses a training method for amplifying training data of a document analysis model. The training method comprises the following steps: acquiring a first sample and a second sample for verification; inputting the original document image into a data generation model based on a generative adversarial network to obtain corresponding pseudo document image data; and obtaining a target sample based on the first sample and a third sample including pseudo document image data, training a document analysis model by using the target sample, and verifying the trained document analysis model by using a second sample, the training of the first model of the data generation model comprises the following steps: carrying out image alignment and image segmentation on a first sample and a corresponding original document image to obtain a plurality of paired image blocks as a first training set; and training the first model by using the first training set to enable the image blocks of the pseudo-legal image genera