A Benchmark Dataset and Evaluation for Non-Lambertian and Uncalibrated Photometric Stereo

Abstract

Classic photometric stereo is often extended to deal with real-world materials and work with unknown lighting conditions for practicability. To quantitatively evaluate non-Lambertian and uncalibrated photometric stereo, a photometric stereo image dataset containing objects of various shapes with complex reflectance properties and high-quality ground truth normals is still missing. In this paper, we introduce the 'DiLiGenT' dataset with calibrated Directional Lightings, objects of General reflectance with different shininess, and 'ground Truth' normals from high-precision laser scanning. We use our dataset to quantitatively evaluate state-of-the-art photometric stereo methods for general materials and unknown lighting conditions, selected from a newly proposed photometric stereo taxonomy emphasizing on non-Lambertian and uncalibrated methods. The dataset and evaluation results are made publicly available, and we hope it can serve as a benchmark platform that inspires future research.

Multi-View Photometric Stereo: A Robust Solution and Benchmark Dataset for Spatially Varying Isotropic Materials

“…Existing multi-view stereo datasets are limited in the variation of lighting conditions and challenging BRDFs, which is the key issue in photometric stereo. Shi et al [44] proposed the 'DiLiGenT' dataset for single-view photometric stereo. We extend it to the multi-view setup and release the 'ground truth' 3D shape as well as camera calibration information to facilitate future research on MVPS.…”

“…At last, to evaluate multi-view photometric stereo algorithms, we build the 'DiLiGenT-MV' dataset, which is a multi-view extension of the 'DiLiGenT' dataset in [44] for benchmarking photometric stereo algorithms. This new dataset contains images of 5 objects of complex BRDFs.…”

Photometric stereo with only two images: A theoretical study and numerical resolution

“…In order to quantitatively assess the accuracy of the proposed PS2 framework, we extracted three images from the dataset [42], which contains 96 images in total, and for which ground truth normals are available. As shown in Fig.…”

Photometric Stereo for 3D Face Reconstruction Using Non Linear Illumination Models

“…Also, worth mentioning the work in [16,9,29] that are based on similar approaches. Recently approaches for 3D reconstruction based on photometric stereo with unknown lighting were proposed [4,25,1,26,31,22,21,27]. Regarding all these methods, the main difference is that they require a significant amount of images captured while the illumination direction is changing, which makes them unsuitable for real time applications and scenarios that involve humans due to their unconstrained self-movement.…”