Author(s):
Yao Yao, Zixin Luo, Shiwei Li, Tian Fang, Long Quan
Description:
Supplimentary result for ECCV 2018 submission. Two differences:
1. Apply per-view depth range selection according to visibilities of sparse points.
2. Apply mean average over depth values during point cloud fusion.
Method parameters:
Same to MVSNet_full (original ECCV submission).
Side-notes:
https://www.cse.ust.hk/~yyaoag/projects.html
Used image set:
Reconstructions from provided image set
Publication:
https://arxiv.org/abs/1804.02505
Software or method website:
https://github.com/YoYo000/MVSNet
Error Visualization:
| Precision | Recall |