In the modern-day social identity is associated with social media and an emoji has foremost significance as well as essential role in the expression of virtual emotions. The main aim of this project is to develop the model that is going to be used in a facial emotion and gender recognition with CNN architecture integration. Then, emotions are going to be sent in a form of avatars. The software is trained on the Kaggle’s ICMP’s 2013-Facial Emotion Recognition (FER) dataset to split it based on gender. Deep learning algorithms is used to real-time classify the facial emotions of a person in a video using the camera. The kind of a person is built up of the face features and the vehemence of the emotions, being depicted in the face expression. Evaluation procedures that are used to verify the system performance and accuracy prove that it is possible for robots to mimic the facial emotions of the person in the video, with different methods. It is a global endeavor that goes beyond departments that are normally associated with entertainment, games, and virtual reality, the latter of which are constantly adhering to the personalized avatar trend. First of all, the proposed system will become a driving force for the growth of the expression of imagination of the people in the virtual reality systems like in the video games world as the immersion and emotional engagement will be increased.