Irregular Shape Image Retargeting Based on Seam Carving

Abstract

Image redirection refers to the size adjustment of digital images to match the aspect ratio of different display terminals. In view of the problem of regular scaling and low computational efficiency of the existing image reorientation methods, the algorithm of irregular-shape image reorientation (Irregular Shape Image Retargeting based on Seam Carving, SC-ISIR) based on joint cropping is proposed. The SC-ISIR algorithm emphasizes image color information through gradient map to avoid image distortion; uses the gradient map as the image importance map for joint cutting, redirect the original image to the minimum rectangular domain that includes the target shape. Finally, select the seam segment at the edge (the Seam Segment Carving, SC) more effectively from the image edge, reconstruct the image into an irregular shape, and improve the effectiveness and calculation efficiency of the algorithm. The experiment verified that SC-ISIR algorithm can transform images into various kinds of irregular shape images.