DEAP: A Database for Emotion Analysis using Physiological Signals

Stimuli selection

The database selected 60 of the 120 stimuli from Last.fm. It allows the users to assign tags to individual songs, thus creating a folksonomy of tags. Meanwhile, Last.fm offers an API, allowing one to retrieve tags and tagged songs. And the rest stimulus were selected manually, with 15 videos selected for each of the quadrants.

Then, for each of the 120 initially selected music videos, a one minute segment for use in the experiment was extracted. In order to extract a segment with maximum emotional content, the author used a linear regression(线性回归) method to calculate arousal and valence for each shot of in movies. The author carried out a equation: ei = sqrt(ai^2 + vi^2) to comput the emotional highlight score of the i-th segment. The ai means arousal and the vi means valence. For each video, the one minute long segment with the highest emtional highlight score was chosen to be extracted for the experiment.

In the end, the final selection of 40 videos used in the experiment was made on the basis of subjective ratings by volunteers.