Emotion Extrication and Analysis on Videos

Video contents are growing at higher rate and facial analysis and emotion detection might be significant for the content providers to publish. This study focuses on extricating the overall emotions conveyed in the video. It is done at two phases: learning phase and the validation phase. The learning phase includes the learning of emotions from human faces using convolutional neural networks and the validation phase includes feature engineering on videos and extrication of emotions conveyed in the video by applying a well learned system. Extensive experiments were conducted on videos and the performance is measured using standard evaluation metrics.