Deep Learning for Age Estimation

A deep learning project done with Dimitri Delpech de Saint Guilhem and Matthieu Souda as part of Polytechnique’s course ‘MAP583 Deep Learning’ by Kevin Scaman.

We studied the topic of Age Estimation, comparing pre-existing models and changing regression models into classification ones. We also studied the impact of fine-tuning some models with augmented dataset (by generating new images with GAN) and ultimately identified potential bias in the original learning datasets.

You can find below the project Slides we’ve done for the project defense.