Would appreciate any form of peer feedback. Very few good MDE models. Trying to emulate how the brain would make use of geometric non-binocular features(occlusion, texture, priors/multi-frame, etc.), and spoon-feed it to a model. Then attempt to generate depth. Will post update once I have results.