Facial recognition is an essential component of human communication, impacting interactions and decision making processes. In order to facilitate more natural and efficient interactions between humans and machines, it is becoming more and more critical to include emotional awareness into machines. In this paper, a Bidirectional Gated Recurrent Unit classification is proposed to quickly recognise the facial reaction. Firstly, image resolution conversion is applied to CK+ facial recognition dataset to resize the image, then pixel values of an image based on its intensity histogram is adjusted using histogram equalization. Then it smoothen the image and removes the noise using Gaussian filter to get high quality image. Next, the processed image is given to feature extraction by Scale Invariant Feature transform it is used to find the key points. After that image features are selected using Grey Wolf Optimization (GWO) designed for further analysis. Finally, a Bidirectional Gated Recurrent Unit (Bi GRU) framework is processed to enhance the classification of face recognition. Using python software proposed framework have accuracy of 92% is accomplished when compared to other techniques.