The big amount of research interest conducted by Facial Emotion Recognition (ER) into the computer vision systems since from last decade due to fact that ER can make the present solutions and products user context aware. There are number of applications like car industry, humanoid robots, internet based services etc. Designing robust and efficient ER method s main research problem with more accuracy and less processing power. Face recognition domain, the first key problem s to detect whether the input face s neutral or having emotion. f it’s detected as emotion, then we have to recognize the type of emotion correctly. n this project, we are proposed novel framework n which first light weight neutral vs. emotion engine of classification to classify the input face to neutral or emotion class. Secondly, f the input face classify as emotional, then we will apply the emotion recognition process. N emotion recognition, we connected the Local Directional Number (LDN) and Feed Forward Neural Network (FFNN) strategies. We utilized the emotion wheels to profoundly distinguish the emotion kind of information face mage. The LDN technique encodes the directional data of the information face surface structure n more packed way. The execution of this approach will be assessed regarding recognition exactness, false discovery rate (FDR), true positive rate (TPR) and handling time utilizing ongoing examination datasets.
Volume 12 | 05-Special Issue
Pages: 13-20
DOI: 10.5373/JARDCS/V12SP5/20201731