No abstract

Research tasks related to human body analysis have been drawing a lot of attention in computer vision area over the last few decades, considering its potential benefits on our day-to-day life. Anthropometry is a field defining physical measures of a human body size, form, and functional capacities. Specifically, the accurate estimation of anthropometric body measurements from visual human body data is one of the challenging problems, where the solution would ease many different areas of applications, including ergonomics, garment manufacturing, etc. This paper formulates a research in the field of deep learning and neural networks, to tackle the challenge of body measurements estimation from various types of visual input data (such as 2D images or 3D point clouds). Also, we deal with the lack of real human data annotated with ground truth body measurements required for training and evaluation, by generating a synthetic dataset of various human body shapes and performing a skeleton-driven annotation.

No abstract

Mesh processing algorithms depend on quick access to the local neighborhood, which requires costly memory queries. Moreover, even having access to the local neighborhood is not enough to efficiently perform many geometry processing algorithms in an automatic or semi‐automatic way. As humans, we often imagine mesh editing at the level of topological information, e.g., altering surface features, adding limbs, etc., which is not supported by current data structures. These limitations come from the widely used mesh representations because the needed information is not implicitly defined by the structure. We propose a novel model representation called Skeletex. Each 3D model is decomposed into two elements: a skeletal structure that encodes the model topology and a vector displacement map to capture fine details of the geometry. Such a co‐representation contains the topology information, as well as the information about the local vertex neighborhood at each texel. Additionally, our data structure facilitates an automatic skeleton‐based cross‐parameterization. This allows us to implement the mesh manipulation tasks in parallel, using a unified streamlined pipeline that directly maps to the GPU. We demonstrate the capabilities of our data structure by implementing surface region transfer and mesh morphing of 3D models.

No abstract