www.icgst.com
home
Password
Community
Styles
Feedback
Sign Up
Sign in
Paper Details:
Downloads:
1293
Serial Number:
P1151539407
Title:
PERFORMANCE ANALYSIS OF SINGLE DICTIONARY LEARNING FOR SINGLE IMAGE SUPER-RESOLUTION
Authors:
Kiran Jadhav and Ramesh Kulkarni and Gaurav Tawde
Abstract:
In this paper, we propose a single dictionary learning algorithm to fully make use of only high-resolution images. Unlike the other methods, dictionary is trained from the set of high-resolution image patches of size 5x5, 7x7 and 9x9 instead of patch pairs of high-/low-resolution images. The advantage of this modification is no need to train the dictionary again when up-scaling is changed. There is a run-time improvement is achieved with best quality of reconstructed image due to single dictionary. The simulation results justify that the proposed method accomplishes the state-of-the-art results compared to other super-resolution methods in terms of both reconstruction ability and with shorter run-time. The demonstration results for single image super-resolution are more promising.
Keywords:
Bicubic Interpolation, Non Local Means, Super-resolution, Sparse Representation, Sparse Coding, Single Dictionary.
Journal/Conference:
International Journal of Graphics, Vision and Image Processing
Volume:
15
Issue:
2
Submission Date:
9/21/2015 12:00:00 AM
Review Date:
10/3/2015 12:00:00 AM
Publishing Date:
10/4/2015 12:00:00 AM
Article Downloads:
1293
Download:
Facebook