Capability of Modified SIFT to Match Stereo Imagery System
This paper presents an improved version of SIFT method for extracting invariant features from images that can be used to solve the correspondence problem between different views of an object or scene in an image. Scale invariant feature transform (SIFT) has recently gained substantial attention in the computer vision community to address the problem. Corresponding features in sequential pairs of images, at various different angular separations, were identified by applying a scale invariant feature transform (SIFT). Due to limitation in the standard SIFT; some of matches are considered false matches. Epipolar-line and disparity window criteria were introduced to enhance the performance of SIFT. Experiments revealed that considerable number of unfaithful matches were removed when new criteria are introduced. Future work will focus on improving the SIFT technique; to rectify the negative matches in order to obtain better matching result.
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.