← sources

Human Pose detection for heel & toe or grounding?

source: reddit (reddit:computervision)published 2026-09-25 09:28ingested 2026-09-26 02:35

retained text

Does anyone use a pose model (or even just point estimation) that detects heels & toes of people? (ankles just aren't useful :)

Or, I feel like I recently saw something that highlighted where a foot touched the ground (an edge)

I'm trying to get my 3D poses to keep contact with the ground better (though I only need good 2D information)

Ultimately I think this is more likely to be solved with some specific coding instead, but curious to see if there's something out there already.

Or, I just start trying to extract masked pixels for feet and given we can assume you never see the bottom of the foot, infer a capsule and use that to intersect with the floor :)

document fetched 2026-09-26 00:23 · canonical https://reddit.com/r/computervision/comments/1wprs6z/human_pose_detection_for_heel_toe_or_grounding