Stereo matching algorithms are useful for estimating a dense depth characteristic of a scene by finding corresponding points from stereo images of the scene. Several factors such as occlusion, noise, and illumination inconsistencies in the scene affect the disparity estimates and make this process challenging. Algorithms developed to overcome these challenges can be broadly categorized as learning-based and non-learning based disparity estimation algorithms. The learning-based approaches are more accurate but computationally expensive. In contrary, non-learning based algorithms are widely used and are computationally efficient algorithms. In this paper, we propose a new stereo matching algorithm using guided image filtering (GIF)-based cost aggregation. The main contribution of our approach is a cost calculation framework which is a hybrid of cross-correlation between stereo-image pairs and scene segmentation (HCS). The performance of our HCS technique was compared with state-ofthe- art techniques using version 3 of the benchmark Middlebury dataset. Our results confirm the effective performance of the HCS technique.
Occlusion is the key and challenging problem in stereo matching, because the results from depth maps are significantly influenced by occlusion regions. In this paper, we propose a method for occlusion and error regions detection and for efficient holefilling based on an energy minimization. First, we implement conventional global stereo matching algorithms to estimate depth information. Exploiting the result from a stereo matching method, we segments the depth map occlusion and error regions into non-occlusion regions. To detect occlusion and error regions, we model an energy function with three constraints such as ordering, uniqueness, and color similarity constraints. After labeling the occlusion and error regions, we optimize an energy function based MRF via dynamic programing. In order to evaluate the performance of our proposed method, we measure the percentages of mismatching pixels (BPR). And we subjectively compare the results of our proposed method with conventional methods. Consequently, the proposed method increases the accuracy of depth estimation, and experimental results show that the proposed method generates more stable depth maps compared to the conventional methods.