herpaderpapotato/pose-vrlens-finetunes-multiclass-v2

17 keypoint pose model finetuned on 180-220deg fov images with 21 keypoints for performer, and 7 keypoints for beholder padded out to 21.

performer keypoints = ['nose', 'left_eye', 'right_eye', 'left_ear', 'right_ear', 'left_shoulder', 'right_shoulder', 'left_elbow', 'right_elbow', 'left_wrist', 'right_wrist', 'left_hip', 'right_hip', 'left_knee', 'right_knee', 'left_ankle', 'right_ankle', 'pelvis', 'umbilicus', 'sternum left', 'sternum right']
beholder keypoints = ['left_hip', 'right_hip', 'left_knee', 'right_knee', 'pelvis', 'umbilicus', 'pelvis_alternate']

not intended to be anatomically correct labels