Predicting Media Memorability Dataset
This dataset is intended to be used for assessing the prediction of how memorable a video will be. The PMMD dataset is a subset of a collection consisting of 1,500 short videos retrieved from TRECVid. Each video consists of a coherent unit in terms of meaning and is associated with two scores of memorability that refer to its probability to be remembered after two different durations of memory retention. Memorability is measured using recognition tests, i.e., through an objective measure, a few minutes after the memorization of the videos (short term), and then 24 to 72 hours later (long term). The dataset consists of 590 videos as part of the training set and 410 additional videos as part of the development set. The videos are shared under Creative Commons licenses that allow their redistribution.
The dataset was validated during the 2020 and 2021 MediaEval Benchmarking Initiative for Multimedia Evaluation.
For more details see:
- De Herrera, A.G.S., Kiziltepe, R.S., Chamberlain, J., Constantin, M.G., Demarty, C.H., Doctor, F., Ionescu, B. and Smeaton, A.F. “Overview of MediaEval 2020 Predicting Media Memorability Task: What Makes a Video Memorable?”, MediaEval 2020.
Acknowledgements:
If you plan to make use of the PMMD dataset, or refer to its results, please acknowledge the work of the authors by citing the paper listed above.
This dataset was conceived using the data gathered during the MediaEval Predicting Media Memorability Tasks. We acknowledge therefore the valuable contribution of the task organizers. Mihai Gabriel Constantin and Bogdan Ionescu's contribution to this work was supported under project AI4Media, A European Excellence Centre for Media, Society and Democracy, H2020 ICT-48-2020, grant #951911.