Infant hungry recognition based on neural network and AR model

Abstract

To deal with nonverbal life was a difficult task. To study their behaviour without knowing what their needs is another crucial issue. A lot of researches have been rapidly investigated. Thus, in this paper we proudly proposed a system to determine the hungry infant based on their facial expression. A Haar Cascade face detection method was implemented. Autoregressive Model (AR) was employed for the coefficient extraction. Some other statistical methods were used as the feature extraction. Finally Neural network (NN) with 93.78% accuracy was accepted.