Snake bite classification using Chain code and K nearest neighbour

Abstract

Live in tropical country causes a high risk of Indonesians affected by snake bites. One of the causes of death due to snake bites is incorrect snake bite identification by visually identifying. The difference between a venomous and non-venomous snake on their anatomy resulting in different bite marks on the victim. Bite marks are useful for medical team for identifying venomous snake bites or not, but there is no system that helps this problem. This research tries to build a system to recognize the existing bite points on the snake bite image and then classified to venomous snake bite or non-venomous using Chain Code and K Nearest Neighbour. The result of recognizing the bite marks has sensitivity value of 75.95%, specificity of 52.7% and accuracy of 65%. Classification of snake bites categories has sensitivity value of 76.92%, specificity value of 85.71% and accuracy of 80%.