Disparity Map Estimation Using Combined Cost Measure

Disparity map generation is necessary in many 3D applications. This paper focuses on disparity map estimation using local based stereo matching. Stereo correspondence algorithms are categorized as global method and local method. Global based methods are computationally expensive. This paper proposes an improved stereo matching algorithm by using local method which combines cost computation of gradient information and histogram of oriented gradient. The results obtained shows better quality disparity map.