Verified, then logged.
Sun 22 Nov 2026 · 00:24 KST / 15:24 UTC한국어 | ENSearch /

Breaking01 / AI

IDWorld identifies people by body features without rendering

The model identifies people by predicting how their features would look in the same outfit, from a fixed angle or as a silhouette.

A monitor in a computational physics workspace shows example simulations on irregular triangular meshes beside a predicted physical field on a different mesh. AI-generated image

Researchers proposed IDWorld on 20 Nov 2026, reporting that the whole-body recognition model ran at 5.9 ms per frame without generating pixels.

The paper describes three conditions for predicting a person's features: the same outfit, a fixed viewing angle and a plain silhouette. The model adjusts the weight it gives each condition for every pair of a query image and a reference image.

In results published by the researchers, IDWorld ran more than 11 times as fast as an approach that rendered pixels directly, with nearly the same accuracy.

The researchers also reported state-of-the-art results on the CCPG and CCGR-MINI benchmarks.

The paper did not specify a date for the code release.