Non-Local versus Bilateral: Multi-adapting disparity map estimation framework

In this work, a new trend in the stereo field is adopted by introducing a hybrid local-global stereo matching techniques. This trend depends on adapting mechanism through a local and global stage. Locally, new adapting similarity measure with new Bilateral and Non-Local adaptive cost aggregation techniques are introduced. Globally, a new energy formulation for the stereo problem is presented in segment domain which tries to make the top benefit from the local stage. Experimental results on the Middlebury dataset demonstrate that our approach is effective in improving the state of the art in the stereo field.