Diagnosis of Burn Images using Template Matching, k-Nearest Neighbor and Artificial Neural Network

The aim of this research is to develop an automated method of determining the severity of skin burn wounds. Towards achieving this aim, a database of skin burn images has been created by collecting images from hospitals, doctors and the Internet. The initial pre-processing involves contrast enhancement in lab color space by taking luminance component. Various pattern analysis or pattern classifier techniques viz. Template Matching (TM), k Nearest Neighbor Classifier (kNN) and Artificial Neural Network (ANN) have been applied on skin burn images and a performance comparison of the three techniques has been made. The help of dermatologists and plastic surgeons has been taken to label the images with skin burn grades and are used to train the classifiers. The algorithms are optimized on pre-labeled images, by fine-tuning the classifier parameters. During the course of research, of the three classifier methods used for classification of burn images it has been observed that the ANN technique reflected the best results. This has been inferred based on the comparative studies of the three methods. In the ANN method the classification of the image of burns has been found to be the nearest to the actual burns. The efficiency of the analysis and classification of the ANN technique has been of the order of 95% for Grade-1 burns, 97.5% for Grade-2 burns and 95% for Grade-3 burns. As compared to 55%, 72.5% and 70% for Grade1, Grade2, and Grade 3 burns respectively for the TM Method and 67.5%, 82.5% and 75% for kNN method. It is therefore felt that the ANN technique could be applied to analyze and classify the severity of burns. This burn analysis technique could be safely used in remote location where specialists’ services are not readily available. The local doctors could use the analyzer and classify the grade of the burn with a good degree of accuracy and certainty. They could start preliminary treatment accordingly, prior to specialists’ services. This would definitely go a long way in mitigating the pain and sufferings of the patients.