COMPUTER ASSISTED ASSESSMENT OF WOUND APPEARANCE USING DIGITAL IMAGING A Hoppe 1 , D Wertheim 1 , J Melhuish 2 , H Morris 2 , K G Harding 2 , R J Williams 3 1 School of Computing, Kingston University, Surrey, UK 2 Wound Healing Research Unit, University of Wales College of Medicine, Cardiff, UK 3 School of Electronics, University of Glamorgan, Pontypridd, UK Abstract – This paper describes a digital image processing system for the analysis of colour in wound images under clinical conditions. The system uses a 3CCD array digital video camera together with a colour scale for reference. The accuracy of colour assessment was compared with clinicians’ assessment of the amount of slough (necrotic tissue) in digital images. The system was found to be within the range of 10% in describing red hue values in red colour patches between different clinical sessions. There was agreement in 75% of the cases between clinicians’ assessment of the amount of slough and a computer assessment using digital imaging. However, colour may not be sufficient to describe wound appearance accurately. I. INTRODUCTION Leg ulcers are often painful and debilitating and about 1.8 per 1000 population develop the condition [1]. Long healing periods cause major costs in treatment for the National Health Service in the United Kingdom. Objective assessment of healing progress is considered important for assessing treatment strategies, however, at present it is difficult to predict how well wounds are healing. A number of parameters have been suggested, such as odour, pain, wound size and appearance of surrounding tissue [2]. The appearance of wounds is considered an important factor in the assessment of healing progress. It has been suggested that analysis of wound colour may be of clinical value [3-6]. However, these studies included only a small number of wounds and either looked at the entire wound or a small section and did not necessarily use colour patches for reference. Some systems were operated under a special experimental setup, which may prove impractical in a day to day clinical routine. II. AIM The aim of this study was to develop and apply a computer system for analysis of wound images under clinical conditions. This includes assessment of colour variability between images taken under clinical conditions with a digital video system. The system would also be used to assess the amount of slough in leg ulcers. III. METHODOLOGY A. Image Acquisition Images of leg ulcers on 30 patients were obtained using a digital video camera (Panasonic NVDX100 B, Matsushita Electric Industrial Co. Ltd., Japan). This camera incorporates three separate CCD arrays, one each for the red, green and blue colour plane. This arrangement provides good colour reproduction and reduces interference between neighbouring pixels [7]. A 10W video light was mounted on the video camera and used to improve lighting conditions. The camera was operated in still picture mode and images were acquired in true-colour and subsequently downloaded digitally to a PC workstation. A graduated scale with additional colour patches (FUJI Colour Scale) was held close to the wound in order to provide a means of colour reference as seen in Figure 1. This was used to compare the variability of colour values between different images. Fig 1: Example of wound image with colour scale