CR-FusionNet is an end-to-end framework that fuses stereo and monocular cues for depth estimation. Using RGB, sparse SGBM depth, and monocular predictions aligned via linear regression, it refines …