3D image denoising using stereo correspondences

In this paper, we propose an image denoising algorithm for 3D demosaicked images. Each color component of the 3D stereo images is processed for noise reduction by identifying the stereo correspondences between two stereo images at subpixel accuracy. Using our proposed method, it has been experimenta...

Full description

Saved in:
Bibliographic Details
Main Authors: Ngai Li, Jimmy Li, Jim S., Randhawa, Sharmil
Format: Conference Proceeding
Language:English
Subjects:
3D images
Algorithms
Arrays
Color
Conferences
demosaicking
Filtering theory
Image color analysis
image denoising
Noise measurement
Noise reduction
Signal to noise ratio
stereo correspondences
Three dimensional
Three-dimensional displays
Visual
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:In this paper, we propose an image denoising algorithm for 3D demosaicked images. Each color component of the 3D stereo images is processed for noise reduction by identifying the stereo correspondences between two stereo images at subpixel accuracy. Using our proposed method, it has been experimentally shown that there is visual improvement in the demosaicked output images using a realistic signal-dependent noise model with demonstration of a higher signal-to-noise ratio (SNR).
ISSN:2159-3442
2159-3450
DOI:10.1109/TENCON.2015.7372926