A New Metric for Measuring Image-Based 3D Reconstruction.

Xu Zhao,Rui Wu,Zhong Zhou,Wei Wu

2012-01-01

Abstract:We propose a novel metric to measure the image-based 3D reconstruction results without ground truth datasets. In contrast to previous metrics, our shading-based metric can not only accurately measure the reconstruction quality but also decouple from any reconstruction algorithm. Considering the uncertainty of topology, texture and soft shadow, we compute an anisotropic irradiance gradient field from multiple images to indicate the regions where reconstruction error occurs. We further apply the metric into the view planning application. Experimental results on both synthetic and real datasets illustrate the effectiveness of evaluating 3D reconstruction by our metric. The reconstruction accuracy and completeness overtop or are the same as the results of manually adding new viewpoints.