Application of genetic algorithms to stereo matching of images
Pattern Recognition Letters - Special issue on genetic algorithms
Constraint, optimization, and hierarchy: reviewing stereoscopic correspondence of complex features
Computer Vision and Image Understanding
Expert Systems: Principles and Programming
Expert Systems: Principles and Programming
A Stereo Matching Algorithm with an Adaptive Window: Theory and Experiment
IEEE Transactions on Pattern Analysis and Machine Intelligence
| Hi-index | 0.00 |
In this paper, the Dempster-Shafer Theory for uncertainty reasoning is presented as a computation tool in designing a model to approach the correspondence problem in Computer Vision. In previous works (Silva and Simoni, 2001a; Silva and Simoni, 2001b) the proposed methodology showed its effectiveness in establishing the correspondence of a pair of images with similar brightness and contrast. In this paper, the efficiency of the uncertainty reasoning methodology is evaluated by applying the method to pairs of real world images with different brightness and contrast. Contextual and structural features of a point are treated as corresponding evidences. The Dempster's rule of combination is used to combine the existing evidences leading to an evidential interval for each candidate point. A search process maximizes the Belief on the combined evidences. The conducted experiments showed the robustness of the approach in establishing the correspondence in situations for which there is illumination and/or focus changes from one real world image to the other.