Apple releases LiTo: It can reconstruct complete 3D objects from a single image, with lighting and shadow effects that are consistent with the viewing angle.

According to 1M AI News , Apple's AI research team published a paper at ICLR 2026 proposing a 3D generation method called LiTo (Surface Light Field Markup), which can generate complete 3D objects from a single image and maintain the consistency of lighting effects such as specular highlights and Fresnel reflections when switching viewpoints.

Previous 3D reconstruction methods could only handle either geometry or diffuse appearance, making it difficult to reproduce lighting and shadow details that change with viewing angle. LiTo encodes both object geometry and view-dependent appearance into a single 3D latent space, and then generates results using a latent flow matching model under single-image conditions. The training data consists of thousands of 3D objects, each rendered from 150 viewpoints and 3 lighting conditions. The decoder learns complete geometry and appearance reconstruction by randomly sampling subsamples. Experiments show that LiTo outperforms the existing method TRELLIS in both visual quality and the accuracy of image reproduction. The paper, authored by Jen-Hao Rick Chang, Xiaoming Zhao (co-first author), Dorian Chan, and Oncel Tuzel, is available on arXiv.