Facial emotion recognition in the wild: an experimental framework and a comprehensive evaluation
Emotion recognition from face images is a challenging task that gained interest in recent years for its applications to business intelligence and social robotics. In real conditions, face images undergo several disturbances that are not extensively covered by esisting in-the-wild benchmarks.
Following on investigations carried out in the field of object recognition, we evaluated the robustness of state-of-the-art models for emotion recognition when their input is subjected to alterations caused by factors present in real-world scenarios.
We constructed two data sets on top of the RAF-DB test set to benchmark the robustness of models for emotion recognition from faces. The two data sets, named RAF-DB-C and RAF-DB-P, contain images modified with 18 types of corruption and 10 types of perturbation, respectively.
The original RAF-DB dataset is available at its own website. For more details of the original dataset, please refer to the paper “Reliable Crowdsourcing and Deep Locality-Preserving Learning for Expression Recognition in the Wild“.
Our corrupted and perturbed dataset can be generated from the original data using the code at this repository. For the sole purpose of allowing our experiments to be reproduced, we offer the possibility to download the data used for our evaluation.
References
If you use these datasets please cite:
- TBD
Download
In order to download the datasets click here.
Help
If you have any problems, do not hesitate to contact us here.