Improvement of Color Images Registration Using CIE Lu’v’Color Space Saturation

Hits: 2034
Research areas:
Year:
2015
Type of Publication:
Article
Keywords:
Saturation of Color Images, Image Registration, Triangle Method, Color Spaces
Authors:
Mojtaba Khaleghi; Mehran Yazdi; Mohammad Zolfaghari
Journal:
IJAIM
Volume:
3
Number:
4
Pages:
165-169
Month:
Jan.-Feb.
ISSN:
2320-5121
Abstract:
One of the most significant tasks when using various sources to integrate and analyze information is image registration. Image registration is a key stage in image synthesis, direction modification and high resolution imaging.In this study, we are going to compare the registration of two input images (reference and rotated) in CIELu'v' color space with N. Nkanza method in two color spaces RGB, HSI. We are going to use image saturation in CIELu'v' color space to implement our suggested algorithm, while protecting image edges. It is quite evident that image saturation means loss of a relatively great number of favorite features of registration (corners). However, the significant rise in speed, along with the higher accuracy of registration, encouraged us to continue the path, especially, about realtime tasks, in which the velocity of image analysis enjoys a special place. Subsequently, finding correlated points of interest, we eliminated unnecessary corner points and, using points of maximum gradient selection technique and applying triangle method, we were released of estimation algorithms for commonly complicated functions of geometric transformation mapping between images. The results of comparing of image registration in the aforesaid color space with color spaces RGB and HSI are proof of this issue.

Indexed By