KBody: Towards general, robust, and aligned monocular whole-body estimation
This addresses the challenge of robust and aligned human body estimation from single images, which is important for applications like computer vision and graphics, though it appears incremental as it builds on existing fitting frameworks.
The paper tackles the problem of monocular whole-body estimation by proposing KBody, a method that fits a low-dimensional body model to images using a predict-and-optimize approach with virtual joints, disentangled optimization, and asymmetric distance fields, achieving improved performance in pose and shape capturing as well as pixel alignment.
KBody is a method for fitting a low-dimensional body model to an image. It follows a predict-and-optimize approach, relying on data-driven model estimates for the constraints that will be used to solve for the body's parameters. Acknowledging the importance of high quality correspondences, it leverages ``virtual joints" to improve fitting performance, disentangles the optimization between the pose and shape parameters, and integrates asymmetric distance fields to strike a balance in terms of pose and shape capturing capacity, as well as pixel alignment. We also show that generative model inversion offers a strong appearance prior that can be used to complete partial human images and used as a building block for generalized and robust monocular body fitting. Project page: https://zokin.github.io/KBody.