A New Method for Image Resizing Algorithm via Object Detection

This article presents a new method to resize the images through object detection that pays attention to higher-levelfeatures of the image. In this method, the offered algorithm identifies the objects in the image by input images,preventing the omission of the identified objects. Afterwards, images are applied to Seam Carving Algorithm to beresized with no change introduced to the selected objects inside the image. In order to learn the algorithm, SVMMethod has been used. Moreover, the article has employed DATA SET, under vocpascal dataset Standard, andDATA SET MIT RETARGETING to learn and test the mentioned algorithm and method respectively. Giving betterresults compared to other methods, this method achieves our goals in keeping the objects inside the image, showingthe images while maintaining high-kevel features in the output.

Keywords:

Object detection , top to bottom feature , considerable object , image resizing , seam carving