In this framework, a feature based dense disparity estimation scheme has been developed. This scheme incorporates segmentation and feature matching schemes for obtaining disparity in different sub-images. The initial features obtained using SURF detector is used for matching segments indirectly. Segments which do not contain any feature point are matched via epipolar constraint by randomly selecting keypoints in the reference image. The scheme is found to provide better texture conservation in the resultant dense image. The proposed algorithm sets a new scheme for computing dense depth from sparse disparity points and requires further improvement in terms of accuracy.
The main.m file needs to be run for obtaining the depth image directly. A readme file has also been provided which can help in providing more details.
Ref: "Poddar, Shashi; Sahu, Hemraj; Bangale, Mohit R; Kumar, Vipan; Kumar, Amod, "Feature based dense disparity estimation," Industrial Instrumentation and Control (ICIC), 2015 International Conference on , vol., no., pp.950,955, 28-30 May 2015
Shashi (2020). Feature based Dense Disparity Estimation (https://www.mathworks.com/matlabcentral/fileexchange/52330-feature-based-dense-disparity-estimation), MATLAB Central File Exchange. Retrieved .