Who Left the Dogs Out? 3D Animal Reconstruction with Expectation Maximization in the Loop
This work addresses the challenge of 3D animal reconstruction from low-quality images for computer vision applications, but it is incremental as it builds on prior methods with a new prior and dataset.
The authors tackled the problem of automatically reconstructing 3D pose and shape of dogs from monocular internet images, achieving results on the Stanford Dog dataset with 20,580 images by learning a richer shape prior through expectation maximization.
We introduce an automatic, end-to-end method for recovering the 3D pose and shape of dogs from monocular internet images. The large variation in shape between dog breeds, significant occlusion and low quality of internet images makes this a challenging problem. We learn a richer prior over shapes than previous work, which helps regularize parameter estimation. We demonstrate results on the Stanford Dog dataset, an 'in the wild' dataset of 20,580 dog images for which we have collected 2D joint and silhouette annotations to split for training and evaluation. In order to capture the large shape variety of dogs, we show that the natural variation in the 2D dataset is enough to learn a detailed 3D prior through expectation maximization (EM). As a by-product of training, we generate a new parameterized model (including limb scaling) SMBLD which we release alongside our new annotation dataset StanfordExtra to the research community.