Removing specular reflection in multispectral dermatological images using blind source separation

---

Authors

Mustapha Zokay¹ , Hicham Saylani¹

1. Laboratoire d’Électronique, Traitement du Signal et Modélisation Physique, Université Ibn Zohr

Conference / event

MIUA: Annual Conference on Medical Image Understanding and Analysis, July 2022 (Cambridge, United Kingdom)

Poster summary

In this poster, we propose a new method for removing specular reflection from multispectral dermatological images which is based on BSS using NMF. The key idea of our method is based on a first step which consists to estimate the number of sources involved, by applying a PCA, then a first version of these sources by applying an ICA to our images. We exploit this first version of the sources in a second step to initialize our NMF algorithm instead of the random initialization that is used by most of the existing methods. In order to quantify numerically the performance of our method, we also propose a new protocol to artificially mix a specular reflection image with a diffuse reflection image. The tests effected on real and artificial multispectral dermatological images have shown the good performance of our method compared to two of the most used methods.

Keywords

Multispectral dermatological images, RGB images, Specular reflection, Diffuse reflection, Blind source separation, Principal component analysis

Research areas

Medical Imaging, Physics, Computer and Information Science

References

1. Yang, Q.,Wang, S., Ahuja, N.: Real-time specular highlight removal using bilateral filtering. In: European conference on computer vision. pp. 87–100. Springer
2. Madooei, A., Drew, M.S.: Detecting specular highlights in dermatological images. In: 2015 IEEE International Conference on Image Processing (ICIP). pp. 4357–4360. IEEE (2015)
3. Lézoray, O. : https://lezoray.users.greyc.fr/researchDatabasesDermoscopy.php

Funding

No data provided

Supplemental files

No data provided

Additional information

Competing interests

No competing interests were disclosed.

Data availability statement

The datasets generated during and / or analyzed during the current study are available elsewhere (e.g., repository).
lezoray.users.greyc.fr/researchDatabasesDermoscopy.php

Creative Commons license

Copyright © 2023 Zokay et al. This is an open access work distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.