Classification couleur pilotée par la segmentation pour l'évaluation de la cicatrisation

We present in this work an automatic classification method applied to the assessment of wound healing. Our approach consists in segmenting the wound into homogenously coloured texture regions to simplify the ulterior classification process. We have tested four unsupervised segmentation methods on our application and the segmentation results have been compared to the ground truth provided by clinicians. The segmented regions were labelled after a supervised learning of a SVM. We obtain a classification rate close to 79.4% during the learning step of SVM over regions of interest provided by the clinicians. Using the JSEG algorithm to segment our images provides overlap accuracy higher than 60%. As we developed a 3D model of the wound, obtained with a simple digital camera by uncalibrated vision techniques, the 2D segmentation results outcome from several views will be mapped onto it to measure true areas.

[1]  Paolo Cignoni,et al.  Derma: Monitoring the Evolution of Skin Lesions with a 3D System , 2003, VMV.

[2]  Thomas Hofmann,et al.  Support vector machine learning for interdependent and structured output spaces , 2004, ICML.

[3]  B. S. Manjunath,et al.  Unsupervised Segmentation of Color-Texture Regions in Images and Video , 2001, IEEE Trans. Pattern Anal. Mach. Intell..

[4]  Benjamin Albouy,et al.  Accurate 3D Structure Measurements from Two Uncalibrated Views , 2006, ACIVS.

[5]  Adilson Gonzaga,et al.  Segmentation and analysis of leg ulcers color images , 2001, Proceedings International Workshop on Medical Imaging and Augmented Reality.

[6]  J. Winder,et al.  New protocol for leg ulcer tissue classification from colour images , 2004, The 26th Annual International Conference of the IEEE Engineering in Medicine and Biology Society.

[7]  W. Westerhof,et al.  Image processing in the study of wound healing. , 1995, Clinics in dermatology.

[8]  A. Hoppe,et al.  Analysis of Skin Wound Images Using Digital Color Image Processing: A Preliminary Communication , 2004, The international journal of lower extremity wounds.

[9]  Peter Plassmann,et al.  Improved active contour models with application to measurement of leg ulcers , 2003, J. Electronic Imaging.

[10]  Lutz Priese,et al.  Fast and Robust Segmentation of Natural Color Scenes , 1998, ACCV.

[11]  P Plassmann,et al.  MAVIS: a non-invasive instrument to measure area and volume of wounds. Measurement of Area and Volume Instrument System. , 1998, Medical engineering & physics.

[12]  P. Plassmann,et al.  An instrument to measure the dimensions of skin wounds , 1995 .

[13]  Daniel P. Huttenlocher,et al.  Efficient Graph-Based Image Segmentation , 2004, International Journal of Computer Vision.

[14]  Vladimir N. Vapnik,et al.  The Nature of Statistical Learning Theory , 2000, Statistics for Engineering and Information Science.

[15]  Huiru Zheng,et al.  Case-based tissue classification for monitoring leg ulcer healing , 2005, 18th IEEE Symposium on Computer-Based Medical Systems (CBMS'05).

[16]  Marina Kolesnik,et al.  Segmentation of wounds in the combined color-texture feature space , 2004, SPIE Medical Imaging.

[17]  Dorin Comaniciu,et al.  Robust analysis of feature spaces: color image segmentation , 1997, Proceedings of IEEE Computer Society Conference on Computer Vision and Pattern Recognition.