Rotation Invariance and Directional Sensitivity: Spherical Harmonics versus Radiomics Features
A. Depeursinge, J. Fageot, V. Andrearczyk, J.P. Ward, M. Unser
Proceedings of the Ninth International Conference on Machine Learning in Medical Imaging (MLMI'18), Granada, Kingdom of Spain, September 16, 2018, [Lecture Notes in Computer Science, vol. 11046, Springer, 2018], pp. 107–115.
We define and investigate the Local Rotation Invariance (LRI) and Directional Sensitivity (DS) of radiomics features. Most of the classical features cannot combine the two properties, which are antagonist in simple designs. We propose texture operators based on spherical harmonic wavelets (SHW) invariants and show that they are both LRI and DS. An experimental comparison of SHW and popular radiomics operators for classifying 3D textures reveals the importance of combining the two properties for optimal pattern characterization.
@INPROCEEDINGS(http://bigwww.epfl.ch/publications/depeursinge1801.html, AUTHOR="Depeursinge, A. and Fageot, J. and Andrearczyk, V. and Ward, J.P. and Unser, M.", TITLE="Rotation Invariance and Directional Sensitivity: {S}pherical Harmonics {\textit{versus}} Radiomics Features", BOOKTITLE="Proceedings of the Ninth International Conference on Machine Learning in Medical Imaging ({MLMI'18})", YEAR="2018", editor="", volume="11046", series="Lecture Notes in Computer Science", pages="107--115", address="Granada, Kingdom of Spain", month="September 16,", organization="", publisher="Springer", note="")