The muscle cells get contracted as part of the SPH simulation. The animation will have to show the same but this falls outside the classic skinning based animation.
Muscles will be stretched in the the direction of muscle bone only. Direction perpendicular to the bone will be ignored.
First proposed solution for muscle animation:
for each muscle cell a bone will be computed - this should be done by computing AABB bounding box of muscle particles
bone space fore each muscle will be computed
particles in each muscle will be transformed from object space into bone tangent space of the muscle, stretched in the bone space along the bone and transformed back to object space
finally, a skinning animation of the cuticle will be applied on the muscle vertices
The muscle cells get contracted as part of the SPH simulation. The animation will have to show the same but this falls outside the classic skinning based animation.