Robust measurement of 3D human skin features via fringe projection profilometry

Abstract

Fringe Projection Profilometry (FPP) is becoming an essential tool in the measurementof 3D human skin features due to its non-contact nature and high precision. Skinfeatures are necessary for monitoring and characterize lesions of the skin. However, theconventional methods for assessing skin lesions are cumbersome and prone to interandintra-observer variations. Recently, FPP 3D scanning technology promises the bestresults in terms of accuracy, fast acquisition, and processing. This thesis contributesto the strengthening of FPP systems in the clinical setting and the problems that arisealong the 3D imaging pipeline. We have dealt with and proposed a solution to theproblems arising in FPP acquisition, especially when measuring human skin. Also,we have explored the main shortcomings of Fourier Transform Profilometry (FTP) forphase retrieval and proposed a modified FTP method to overcome these limitations.Furthermore, we have introduced a Simultaneous Phase Unwrapping and Denoisingalgorithm for phase imaging in challenging conditions. Finally, we have described aframework for the robust automated measurement of 3D skin features from the SkinPrick Test. Through experimental results, we have shown that the proposed methodsprovide accurate and reliable measurements of human skin features.