METHOD FOR REGION EXTRACTION, METHOD FOR MODEL TRAINING, AND DEVICES THEREOF

摘要

Method for region extraction, method for model training, and devices thereof are provided in the disclosure, which belongs to image processing field. The method for region extraction may include: acquiring (201) a recognition model, the recognition model being formed by training predefined sample images with a classification algorithm, the predefined sample images including positive sample images and negative sample images, each of the positive sample images having at least one numeral character, each of the negative sample images having no numeral character or incomplete numeral characters only; performing (202) recognition on an image to be recognized, in accordance with the recognition model, to obtain at least one numeral region; and performing (203) region segmentation on the numeral region to obtain at least one single-numeral region.