Evaluation of Wound Healing Process Based on Texture Analysis Christos P. Loizou 1,2 , Takis Kasparis 2 , Orthodoxia Mitsi 2 , Michalis Polyviou 3 1 Department of Computer Science, Intercollege, Limassol, Cyprus, Email: loizou.c@lim.intercollege.a.c.cy 2 Department of Electrical Engineering, Computer Engineering & Informatics, Cyprus University of Technology, Limassol, Cyprus, Email: christos.loizou@cut.ac.cy; takis.kasparis@cut.ac.cy 3 Polyviou Clinic, Limassol, Cyprus, Email: michalis_polyviou@yahoo.co.uk Abstract—Wound healing rate, remains an interesting and important issue, in which modern imaging techniques have not yet given a definitive answer. In order to guide better therapeutic interventions, a better understanding of the fundamental mechanisms driving tissue repair are required. The wound healing rate is primarily quantified by the rate of change of the wound’s surface area. The objective of this study was to establish a standardised and objective technique to asses the progress of wound healing in foot by means of texture analysis. The methods of image pre-processing, segmentation and texture analysis together with visual expert’s evaluation were used to assess the wound healing process. A total of 40 digital images from ten different subjects with food wounds were taken every third day, for 12 days, by an inexpensive digital camera under variable lighting conditions. The images were intensity normalized, and wounds were automatic segmented using a snake’s segmentation system. From the segmented wounds 15 different texture characteristics and 4 different geometrical features were extracted in order to identify features that quantify the rate of wound healing. We found texture characteristics that may indicate the progression of wound healing process. More specifically, some texture features increase (mean, contrast), while some other texture features decrease (entropy, sum of squares variance, sum average, sum variance) with the progression of the wound healing process. Some of these features were found to be significantly different in a specific time point and this could be used to indicate the rate of wound healing. No significant differences were found for all geometrical measures. The results of this study suggest that some texture features might be used to monitor the wound healing process, thus reducing the workload of experts, provide standardization, reduce costs, and improve the quality for patients. The simplicity of the method also suggests that it may be a valuable tool in clinical wound evaluation. Future work will incorporate additional texture and geometrical features for assessing the wound healing process in order to be used in the real clinical praxis. Keywords-Wound segmentation; texture analysis; wound healing rate. I. INTRODUCTION Chronic wounds present an increasing health challenge as the population ages and the incidence of different chronic diseases grows worldwide [1]. The progress of wound healing may be quantified by the rate of change of the wound’s surface area [2]. However, this is a challenging task due to the complexity of the wound, the variable lighting conditions, and the time constraints in clinical laboratories. A color image of a wound on foot is presented in Fig. 1a). One way to evaluate wound healing rate is to monitor wound status by taking images of the wound at regular patient visits (see Fig. 2). If the physical dimensions of the wound are assessed at regular time intervals, then the experts will know if the patient is responding well or not to a particular treatment and if necessary change it [2]. In 2012 the 22 nd annual meeting of the Wound Healing Society (WHS), set the standards for wound healing procedures and proposed recommendations for evaluating the optimal wound treatment [3]. There are not many research groups worldwide that are involved in color image processing of wounds. In [2], the authors proposed and evaluated an algorithm for the wound segmentation with minimal manual input and a high accuracy, which uses a combination of both RGB and L*a*b* color spaces, as well as a combination of threshold and pixel-based color comparing segmentation methods. Jones et al. [4], and Jones [5], developed the MAVIS system, which is able to automatically measure the dimensions of skin wounds. Their method was based on color segmentation algorithms and was able to segment an image into one of three tissue types: healthy skin, wound tissue and epithelialisation tissue. Furthermore, six measurement parameters: the R, G and B color planes, hue, saturation and gray-level intensity were taken into consideration. The R, G and B color planes were only examined in isolation showing that straightforward thresholding of color planes cannot produce a good segmentation which distinguishes between wound and skin tissues. They found that wound segmentation is only partially succeeded, if only the 1D color histograms were taken into consideration, while using a 3D RGB histogram space, the color volume clusters may be more widely separated and a better segmentation result can be achieved. Mekkes et al. [6], made some progress with such the 3D RGB color histogram clustering technique to asses the healing of wounds. It was shown that clusters in RGB space for a given tissue type formed an irregularly shaped 3D cloud, and so simple thresholding along the R, G and B axes would not help to segment the image into these three tissue types. Some other researchers presented their techniques on the segmentation of wounds in color mages based on the use of the black-yellow- red classification scheme to evaluate the debridement activity of wounds [7]. A method to correct for limb convexity in color video images in order to measure the size of skin wounds and Proceedings of the 2012 IEEE 12th International Conference on Bioinformatics & Bioengineering (BIBE), Larnaca, Cyprus, 11-13 November 2012 978-1-4673-4358-9/12/$31.00 ©2012 IEEE 709