Indoor scene understanding is central to applications such as robot
navigation and human companion assistance. Over the last years, data-driven
deep neural networks have outperformed many traditional approaches thanks to
their representation learning capabilities. One of the bottlenecks in training
for better representations is the amount of available per-pixel ground truth
data that is required for core scene understanding tasks such as semantic
segmentation, normal prediction, and object edge detection. To address this
problem, a number of works proposed using synthetic data. However, a systematic
study of how such synthetic data is generated is missing. In this work, we
introduce a large-scale synthetic dataset with 400K physically-based rendered
images from 45K realistic 3D indoor scenes. We study the effects of rendering
methods and scene lighting on training for three computer vision tasks: surface
normal prediction, semantic segmentation, and object boundary detection. This
study provides insights into the best practices for training with synthetic
data (more realistic rendering is worth it) and shows that pretraining with our
new synthetic dataset can improve results beyond the current state of the art
on all three tasks.
Yinda Zhang, Shuran Song, Ersin Yumer, Manolis Savva, Joon-Young Lee, Hailin Jin, and Thomas Funkhouser.
"Physically-Based Rendering for Indoor Scene Understanding Using Convolutional Neural Networks."
IEEE Conference on Computer Vision and Pattern Recognition (CVPR), July 2017.
author = "Yinda Zhang and Shuran Song and Ersin Yumer and Manolis Savva and
Joon-Young Lee and Hailin Jin and Thomas Funkhouser",
title = "Physically-Based Rendering for Indoor Scene Understanding Using
Convolutional Neural Networks",
booktitle = "IEEE Conference on Computer Vision and Pattern Recognition (CVPR)",
year = "2017",
month = jul