Bilateral Filtering Using New Histogram and Nearest Neighbor Searching

S.S. Waghchaware, Dr. Chanakya Kumar Jha and S.R. Dhanokar


Abstract:

Noise is main factor that deteriorates image quality which is due to random changes in color and brightness information. Noise is classified into different categories such as Gaussian, impulse, mixed noise etc. In this paper, we have used locality sensitive histogram (LSH) in two applications for bilateral filtering and for nearest neighbor field (NNF). In first application, LSH is extended to bilateral filtering to preserve edges in images. In second application, NNF is computed, in which all pixels in other image are mapped to find out most analogous patch in other image. Bilateral filter uses LSH to avoid drawback of box filter. The new histogram is as efficient as that of LSH which gives better accuracy. Also NNF can be estimated which is not dependent of patch size.

Issue: 08-Special Issue

Year: 2018

Pages: 1564-1569

Purchase this Article