Document image layout understanding method and device and electronic equipment

• Main Authors: WANG YAMENG, LIU ZHENGZHEN
• Format: Patent
• Language: Chinese; English
• Published: 12.09.2023
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; PHYSICS

The invention discloses a document image layout understanding method and device, and belongs to the technical field of image processing. The method comprises the steps of performing feature extraction on a to-be-processed document image through a backbone network of a pre-trained layout understanding model to obtain image features; processing the image features through a first branch network of a layout understanding model to obtain a first detection result of coarse-grained layout elements included in the to-be-processed document image, and processing the image features through a second branch network of the layout understanding model to obtain a second detection result of coarse-grained layout elements included in the to-be-processed document image; obtaining a second detection result of fine-grained layout elements included in the to-be-processed document image; and according to the first detection result and the second detection result, executing a detection result joint optimization operation to obtain a