Overview | References | Data | Presentation | About us
This work proposes a series of Deep Learning models aimed at addressing the problem of recognizing emotions through facial expressions. The experiments were conducted on a dataset made by us, aggregating images acquired via webcam and wild images collected by Google, augmented with Data Augmentation techniques.
For the best model, resulting in a convolutional network associated with a Feature Extraction by VGGFace, several architectures were explored, followed by a pipe of Hyper Parameter Optimization (HPO) with Sequential Model-Based Optimization (SMBO) methods.
[1] P. Abhang, S. Rao, B. W. Gawali, and P. Rokade (2011), Emotion recognition using speech and eeg signal a review, International Journal of Computer Applications, vol. 15, pp. 37–40
[2] P. Ekman (1971), Universals and cultural differences in facial expressions of emotion, Lincoln University of Nebraska Press, Nebraska, USA
[3] A. Kołakowska, A. Landowska, M. Szwoch, W. Szwoch, and M. R. Wróbel (2014), Emotion Recognition and Its Applications, Cham: Springer International Publishing, pp. 51–62
[4] Q. Cao, L. Shen, W. Xie, O. M. Parkhi, A. Zisserman (2018), VGGFace2: A dataset for recognising face across pose and age, International Conference on Automatic Face and Gesture Recognition
The dataset used was obtained thanks to the union of three different data sources, in which the same data are found in the form of images of different dimensions.
The data used for this project are:
- RECData » available in the RECData folder (soon);
- VISGRAF » available at the following link;
- Scraping » available in the Google data folder (soon).
Our slides presentation is available in the slides folder.
Here we show only the cover:
⊜ Alessandro Borroni
- Current Studies: Data Science M.Sc. Student at Università degli Studi di Milano-Bicocca (UniMiB);
- Background: Bachelor degree in Business Economics at Università degli Studi di Milano-Bicocca (UniMiB).
⊜ Andrea Corvaglia
- Current Studies: Data Science M.Sc. Student at Università degli Studi di Milano-Bicocca (UniMiB);
- Background: Bachelor degree in Physics at Università degli Studi di Milano-Bicocca (UniMiB).
⊜ Massimiliano Perletti
- Current Studies: Data Science M.Sc. Student at Università degli Studi di Milano-Bicocca (UniMiB);
- Background: Bachelor degree in Ingegneria dei materiali e delle nano-tecnologie at Politecnico di Milano.