Skin Burn Detection using Feature Extraction

Abstract

Feature extraction is one of the most interesting approaches to the identification of any application area. Nowadays medical science is using feature extraction in every medical imaging field as well as in the various medical science areas. Similarly, the Burn is also one of the critical disease which requires fast diagnosis to safe the patient. In such cases, the diagnosis is more important, so that the proper cure can be done. The new method using CNN and augmentation is one of the best approaches in this area which provides more correct results. The proposed model is basically for the identification of burn area as well as the impact of burn on the body part. The results are compared after applying the augmentation technique, so that the training can be done on more number images, As the BIS data set contains only 90 images, therefore there is need to have more images for the evaluation of the performance of the proposed approach. The model provides 94% accuracy in the training set and 92% in the testing set. This paper presents the extension work of the CNN model using augmentation to improve the outcome.