Published Paper Details:

Artificial intelligence has its demand in almost every possible field, such as, medical, agriculture, military and so on. In depth, it is used to detect cancereous cells, lane detection, natural language processing, number plate recognition, and also in identification of fire. Fire has its own set of behavioural traits. Its static behaviour includes colour, texture, brightness and edges. characteristics for identification. The movement characteristics, shape which fluctuates incessantly constitute the dynamic behaviour of fire. The proposed approach presents an effective way of recognizing fire. It consists of two stages namely, the dynamic behaviour extraction and static behaviour extraction followed by classification for which pre-trained neural network is used to achieve accuracy. This termed as transfer learning. The first stage aids in ruling out all the unnecessary characteristics. The output of the first stage goes through the second stage, where all its static features are studied to put it under the correct category of classification.