Great article! Just a question regarding monocular depth estimation. So in stereo, we first find disparity then compute for depth using the defined equation, but does the monocular depth estimation neural network directly spit out depth values (m) for each pixel?