The prediction of one’s emotional response to affective movie content has been a challenging task in affective computing. Previously, work has been done on the same in regards to Hollywood movies. However, the great diversity of Bollywood movies is often missed out from such projects. This paper will take into account specifically Bollywood content. Firstly, a database of Bollywood movie clips will be created by collection and annotation of perceived emotions via crowdsourcing. We will then collect data for induced emotions by using facial expression recognition tools. Along with facial expression, we will also collect data of EEG (Electroencephalogram) signals and GSR (Galvanic Skin Response) signals. Lastly, with the application of machine learning a model will be trained on half of the data recorded to predict the emotions of movie audiences in another half of the data.
: Aarushii Agrawal (IIIT Delhi), Shahid Nawaz Khan (IIIT Delhi)