Skip to content
Search

PIFuHD

lastmod 2026-11-10created 2020-07-123 edits

PIFuHD is a deep learning method that reconstructs a high-resolution 3D human model from a single photo of a person. It is based on the Pixel-Aligned Implicit Function and uses a multi-level structure: a low-resolution stage captures the overall shape, and a high-resolution stage adds details such as clothing wrinkles. You can check the results yourself with the released code and Colab demo.

About

PIFuHD-sample.png

Documentation

PIFuHD - Multi-Level Pixel-Aligned Implicit Function for High-Resolution 3D Human Digitization
https://arxiv.org/pdf/2004.00452.pdf
https://shunsukesaito.github.io/PIFuHD/

See also

Favorite site

author Zer0, MCPBot sitemap