Table 7.

Generalization performance of the proposed method on additional datasets

Dataset	Avg Error (pixels)	rmse (pixels)	Comments
Middlebury-v3 [53]	12.345	5.678	High quality stereo images
KITTI-15 [54]	16.789	7.234	Real-world driving scenarios
NYU Depth Dataset V2 [55]	19.980	8.456	Indoor scenes with varied lighting