Extract body parts in addition to joints in human pose estimation API

I am using the human pose estimation API and I am wondering if the API calculates the pixels associated with a body part prior to calculating the joint location? That is, for every pixel in the image, what is the likelihood that this pixel is part of the right bicep, for one example? If that is being calculated, it would be a great help to me to expose that information in the poseResult structure. Or if a helper function could access that information. I am assuming that this information is being calculated as Part Affinity Fields (in some way similarly to Open Pose) as part of the joint location determination.

Unfortunately, that information isn’t available for our pose models. We’ll make a note of the feature request and reach out if it is added to the API in the future.

Thanks Jameson, I appreciate the quick response and feature request addition.