Search Results for author: Yuto Enyo

Found 2 papers, 0 papers with code

Diffusion Reflectance Map: Single-Image Stochastic Inverse Rendering of Illumination and Reflectance

no code implementations7 Dec 2023 Yuto Enyo, Ko Nishino

In this paper, we introduce the first stochastic inverse rendering method, which recovers the attenuated frequency spectrum of an illumination jointly with the reflectance of an object of known geometry from a single image.

Inverse Rendering

nLMVS-Net: Deep Non-Lambertian Multi-View Stereo

no code implementations25 Jul 2022 Kohei Yamashita, Yuto Enyo, Shohei Nobuhara, Ko Nishino

Our key idea is to formulate MVS as an end-to-end learnable network, which we refer to as nLMVS-Net, that seamlessly integrates radiometric cues to leverage surface normals as view-independent surface features for learned cost volume construction and filtering.

Cannot find the paper you are looking for? You can Submit a new open access paper.