IMAGE TABLE EXTRACTION METHOD, APPARATUS, ELECTRONIC DEVICE AND STORAGE MEDIUM

摘要

The embodiment of the present application discloses a method of extracting a table of images, an apparatus, an electronic device, a storage medium, and a training method of a table extracting model, and relates to the field of artificial intelligence and cloud computing technology, and the step of acquiring an image to be processed, a table Based on the extraction model, a table of images to be processed is generated, wherein the table extraction model is obtained based on field position characteristics, image characteristics, and text characteristics of the sample image, and text information of the image to be processed is table And, on the one hand, through the field location feature, the relationship of each field on the location can be determined, and it can be applied to an image that is at least a partially dotted box table, and the flexibility and versatility of table recognition Implement technology effects that can improve the performance; On the other hand, by taking full consideration of field location features, image features, and text features, that is, by analyzing images to be processed in various dimensions, features of various dimensions are obtained, thereby improving the accuracy and reliability of table recognition. Implement.