PROJECT TITLE :

A Novel Image Representation via Local Frequency Analysis for Illumination Invariant Stereo Matching - 2015

ABSTRACT:

During this paper, we have a tendency to propose a unique image representation approach to tackle illumination variations in stereo matching issues. Images are mapped using their Fourier transforms which are convolved with a collection of monogenic filters. Frequency analysis is allotted at different scales to account for many image content. The part congruency and therefore the native weighted mean phase angle are then computed over all the scales. The original image is transformed into a replacement illustration using these two mappings. This illustration is invariant to illumination and distinction variations. More importantly, it's generic and will be used with most sparse and dense stereo matching algorithms. Likewise, sequential feature matching or tracking can conjointly benefit from our approach in varying radiometric conditions. We have a tendency to demonstrate the improvements introduced with our image mappings on well-established data sets in the literature along with on our own experimental scenarios that include high dynamic range imagery. The experiments embody each dense and sparse stereo and sequential matching algorithms where the latter is considered within the very challenging visual odometry framework.