Fast stereo matching with fuzzy correlation

Stereo matching is an extensively researched topic in computer vision. Stereo matching algorithms are essential for recovering depth information of objects. Existing state-of-the-art stereo methods require very high processing times. Consequently, we cannot employ them in commercial applications though they are very accurate and robust. With a view to reduce the computation time this paper presents a fast and efficient stereo algorithm using fuzzy correlation measure which is the core contribution of this paper. We attempt to lessen the computation time significantly while keeping the accuracy at a reasonable stage so that the proposed algorithm can be employed in real time applications. Experimental evaluation proves the robustness and effectiveness of our proposed method comparable to other methods in terms of computational efficiency and accuracy.