Home > Published Issues > 2015 > Volume 3, No. 1, June 2015 >

Enhanced Depth Features of Human Bodies for Image Retargeting

Jen-Jung Cheng 1, Chin-Chen Chang 1, and Huei-Yung Lin 2
1. Department of Computer Science and Information Engineering, National United University, Miaoli 360, Taiwan
2. Department of Electrical Engineering and Advanced Institute of Manufacturing with High-tech Innovations, National Chung Cheng University, Chiayi 621, Taiwan

Abstract—In this paper, we propose enhanced depth features of human bodies for image retargeting. First, the depth map of an input image is generated. Second, with the depth map we divide the depths into several depth layers to separate human bodies and the background. Next, the proposed approach performs the morphological dilation operation on the human bodies and enhances the edges of the human bodies to generate an enhanced depth map. Moreover, we assign different weights for the human bodies and their edges according to their importance. Finally, we generate the target image based on the enhanced depth map. The proposed approach can obtain pleasing resized images for a wide range of images.

Index Terms—image retargeting, feature, depth map, seam carving

Cite: Jen-Jung Cheng, Chin-Chen Chang, and Huei-Yung Lin, "Enhanced Depth Features of Human Bodies for Image Retargeting," Journal of Image and Graphics, Vol. 3, No. 1, pp. 52-55, June 2015. doi: 10.18178/joig.3.1.52-55