Search

There is an ultimate goal to design a system where machine can recognize natural interaction between humans and machines. It’s crucial to design such system based on the HCI systems. This field is widely studied by many Computer Vision (CV) and Machine Learning researchers. Even many systems, based on the study has been designed and implemented on the basis of its popularity such as Gesture recognition based HCI application, FER systems and the combination of these two to get better output in a human emotion recognition system. In the visual representation of gesture recognition such as body, hand, face, sign, emotions, expressions, movements of body parts are used to build such interactive system. Many challenges occurred during the design such as lightning condition, background, head orientation, hand orientation, location, position of body gesture, static and dynamic movements and, so on. FER based on the hand gesture is one of the challenging and main research highlights in the field of HCI. Through facial expressions, we can convey the useful, natural, intension to the other person during the communication. Also, with the hand gesture and its position during the communication, gives more accuracy in the recognition process.

The face and hand is the main part of the human body for interaction during communication. Human emotions through facial and hand gesture is the result of the muscle movement of eyes, nose, head, lips, and hand gesture. The importance of this recognition system is beneficial to many different real-time fields such as psychology, neurology, education, and medical science for the patients specifically as well as the day-to-day life like driver monitoring systems, social interactions, cognitive science, automated tutoring systems, or smart environments which are mainly part of the Human–Computer Interaction (HCI). The aim of this work is to analyze the recent advances in image processing and machine learning techniques with respect to facial emotion recognition based on hand position surrounding the face through the hand gesture. Here, PCA is used for feature extraction and for the normalization Dynamic Time Warping (DTW) is used. Viola–Jones Methods are used to identify prominent features of the face and hand; Gabor filters are used to represent facial geometry at selected locations of fiducial points and combination of multi-class AdaBoost with feature vector and the combination of Hidden Markov Model (HMM) and Support Vector Machines (SVM) is used for the classification of the human emotions based on hand gesture.