Methods, devices, and computer readable storage media for image processing

Abstract

Embodiments of the present disclosure relate to methods, devices and computer-readable storage medium for image processing. A method for image processing comprises obtaining a plurality of images, each of the plurality of images having an initial semantic segmentation label indicating a semantic class of a pixel in the each image; obtaining a plurality of image masks corresponding to the plurality of images, each of the plurality of image masks being used for selecting a target region in a corresponding image of the plurality of images; regenerating respective semantic segmentation labels of the plurality of images based on the plurality of image masks and initial semantic segmentation labels of the plurality of images; and generating a mixed image and a semantic segmentation label of the mixed image based on the plurality of images and the regenerated respective semantic segmentation labels. By using the generated mixed image and its semantic segmentation label as training data to train an image semantic segmentation model, it helps to reduce the redundant learning on easy training samples for the model and mitigate the model overfitting problem, thereby enhancing the model performance.