Towards Automatic Acne Detection Using a MRF Model with Chromophore Descriptors

Zhao Liu 1, * Josiane Zerubia 1
* Corresponding author
Abstract : This paper proposes a new acne detection approach using a Markov random field (MRF) model and chromophore descriptors extracted by bilateral decomposition. Compared to most existing acne segmentation methods, the proposed algorithm enables to cope with large-dynamic-range intensity usually existing in conventional RGB acne images captured under uncontrolled environment. Algorithm performance has been tested on acne images of human face from a free public database. Experimental results show that acne segmentation derived from this new approach highly agrees to human visual inspection. Moreover, inflammatory response and hyperpigmentation scar can be well discriminated. It is expected that a computer-assisted diagnostic system for acne severity evaluation will be constructed as a consequence of the present work.
Liste complète des métadonnées

Cited literature [6 references]  Display  Hide  Download
Contributor : Zhao Liu <>
Submitted on : Wednesday, September 18, 2013 - 10:50:13 AM
Last modification on : Saturday, January 27, 2018 - 1:31:40 AM
Document(s) archivé(s) le : Friday, December 20, 2013 - 2:37:12 PM


Files produced by the author(s)


  • HAL Id : hal-00863046, version 1



Zhao Liu, Josiane Zerubia. Towards Automatic Acne Detection Using a MRF Model with Chromophore Descriptors. European Signal Processing Conference (EUSIPCO), Sep 2013, Marrakech, Morocco. ⟨hal-00863046⟩



Record views


Files downloads