Face Emotion Recognition for Crowd is a software tool (written in Python) that enables the analysis of facial expressions in a video.

In FER4Crowd you can set different settings (time frame, frame resolution, sampling interval, etc.) and select among several face emotion classifiers.

An offline or Google Drive report is generated at the end of the video processing. It shows logs, charts and statistics, useful for a complete and deep affective analysis. Moreover, FER4Crowd provides a heuristic model to evaluate the performance of different FER (Face Expression Recognition) systems on the video which is being analyzed; it samples a certain number of frames and generates manual annotation modules, which will then be compared with the results of the classifiers.

The requirements guiding the FER4Crowd development emerged within the DoppioGioco project aimed at investigating the emotional response of audience and actors during a theatrical performance.

FER4Crowd has been developed by Antonio D’Ambrosio for his Laurea Magistrate final Dissertation, under the supervision of Rossana Damiano.