ABSTRACT
Human facial expression recognition plays a mesmerizing role in each and every field of technological advancement. As our world, every time moves one step closer to technological advancement the dependency of humans on technology rapidly shoots ups. Generally, facial expression is a direct and phenomenal way of protecting human feelings for one another. Then what might be the outcome if the humans are understood by the machines? Here propose FER detection and extraction via a Histogram of Oriented Gradient (HOG). For starters, HOG upholds methods that can improve the accuracy and robustness of the algorithm used. Secondly, six universal emotions are detected by using lightweight Convolutional Neural Networks (CNN). This framework is used for the detection and transmission of the face coordinates to the facial expressions classification with additionally employing Multi-task Cascaded Convolutional Networks (MTCNN). This is mainly applicable in the medical research field, security, customized marketing, and augmented reality.
FER, Multi-task convolutional neural network (MTCNN), Histogram of oriented gradient (HOG)
