SMALR means "SMAL with Refinement". To estimate animal 3D shape and texture we exploit the SMAL model to estimate pose and provide an initial shape estimation. Then, in the refinement step we deform the estimated initial shape to better match image evidence.
Here is the SMAL model we have used for initialization and the CVPR 2018 paper results. By downloading our data you agree to the license terms.
If you are unsure whether your use counts as as non-commercial please contact firstname.lastname@example.org before downloading.