Segmentation Of Wound By K-Means Clustering And Automatic Prediction Of Healing Time

Abstract

A wound is any damage or break on the surface of the skin. Wounds can be injuries caused by burns, abrasions, skin tears, and surgical. They also occur because of underlying diseases. All these do not have the same healing time as it depends on the type of wound, its initial reaction, and the time taken to form new skin. Wounds generally heal within one to eight weeks depending on its nature. Non-healing types that do not heal within this period. A non-invasive type of wound assessment is proposed in this study. Wound images were collected and are color segmented by K-means clustering method. Then the parameters such as area, perimeter, length, and width are determined automatically by the code written using MATLAB software. Using these parameters an automatic prediction of healing time is estimated. The user had to load the image of the wound into the developed software and the prediction of healing in given terms of days. This has the potential to be developed as a mobile application so that it can be used by the health care professionals and laymen. This type of prediction reduces the anxiety of the suffering.