Skip to main content

Table 2 Qualitative comparison of photometric stereo algorithms

From: Robust surface normal estimation via greedy sparse regression

 

Robustness

Smoothness

Efficiency

 

Overall

Highlight

Shadow

Concavity

  

LS

Very low

Very low

Very low

Very low

Very high

Very high

LMS

Very high

Very high

Very high

Low

Very low

LMS mode

Low

Low

Low

Medium

Very high

IRL1

High

High

High

Medium

High

High

SBL

High

High

High

Medium

High

Low

OMP

Very high

Very high

Very high

Low

Medium

High

  1. The performance is evaluated on a five-level scale: ‘Very low’, ‘low’, ‘medium’, ‘high’ and ‘very high’. Fields that are not available are indicated by a ‘–’ sign