I saw the documentary Montage of Heck and really enjoyed the animated segments. I've been able to find a few articles describing a vague depiction of how Kurt's doodles were compiled and composited, however still not sure about image prep and character rigging after the initial scan.
Were the images simply cropped using the pen too (or something similar)l? Vectorization?
The images were cut into pieces in Photoshop, limbs separated out onto their own layers, missing pieces filled in, then brought back into AE as a Comp (cropped layers). Anchor points were placed in rotation points (shoulders, hips) and parented to torso and the rest (elbow, wrist, knee, ankle etc. joints) animated with puppet pin tools. Puppet pins need nulls parented to all parents up to torso.
That's what it looks like.
Or you could cut the characters into smaller pieces (at elbows, wrists, knees and ankles) and do the rigging with DUIK. Doesn't look like this was done that way, but it would be close to the same result.