MeTTA: Single-View to 3D Textured Mesh Reconstruction with Test-Time Adaptation

Kim Yu-Ji1, Hyunwoo Ha2, Kim Youwang2, Jaeheung Surh3, Hyowon Ha3,†, Tae-Hyun Oh1,2,4,†
1Grad. School of AI, POSTECH, 2Dept. of Electrical Engineering, POSTECH, 3Bucketplace, 4Institute for Convergence Research and Education in Advanced Technology, Yonsei University
†denotes corresponding authors
British Machine Vision Conference (BMVC) 2024
arXiv GitHub Poster

We propose MeTTA, which closes the domain gap between training and test time by jointly updating mesh, texture, and viewpoint with a carefully designed learnable virtual camera by leveraging the generative model prior.

Abstract

Reconstructing 3D from a single view image is a long-standing challenge. One of the popular approaches to tackle this problem is learning-based methods, but dealing with the test cases unfamiliar with training data (Out-of-distribution; OoD) introduces an additional challenge. To adapt for unseen samples in test time, we propose MeTTA, a test-time adaptation (TTA) exploiting generative prior. We design joint optimization of 3D geometry, appearance, and pose to handle OoD cases. However, the alignment between the reference image and the 3D shape via the estimated viewpoint could be erroneous, which leads to ambiguity. To address this ambiguity, we carefully design learnable virtual cameras and their self-calibration. In our experiments, we demonstrate that MeTTA effectively deals with OoD scenarios at failure cases of existing learning-based 3D reconstruction models and enables obtaining a realistic appearance with physically based rendering (PBR) textures.

MeTTA

We propose a test-time adaptation pipeline to reconstruct a 3D mesh with PBR texture from a single-view image. “Ref. Image” refers to the reference input image. “Seg. Image” refers to the object-segmented image from “Ref. Image”.

Cross-domain Results

In-domain Results (Pix3D)

PBR Editing

-->