Survey of Emotion Detection Based on Text and Facial Modalities

Humans express their feelings directly or indirectly through their facial expression, speech, writing or gestures. In the current era, people express their feelings through social media, news, articles and micro blogs. It means of emoting varies from location, culture, gender etc. Due to advancement computer robotics, new angles in human-computer interactions have been brought about which are affecting our day to day life. Real time face to face communication requires quick and accurate assessments to make computers comprehend human needs and improve on its ability to communicate. An important basis for these assessments is human emotion. To tackle emotion detection problem more precisely, researchers all around the world are finding effective ways through text, speech, psychology and facial modality. This Survey covers few of the already existing emotion recognition model datasets, their techniques and features. The paper focuses on reviewing emotions based on text and facial modalities. It recaps the achievements in the field of emotion detection and has highlighted extensions for better outcome.