Beyond Accuracy: Fairness, Scalability, and Uncertainty Considerations in Facial Emotion Recognition

Laurits Fromberg, Troels Nielsen, Flavia Dalia Frumosu, Line H Clemmensen
Proceedings of the 5th Northern Lights Deep Learning Conference ({NLDL}), PMLR 233:67-74, 2024.

Abstract

Facial emotion recognition (FER) from images or videos is an emerging subfield of emotion recognition that in recent years has achieved increased traction resulting in a wide range of models, datasets, and applications. Benchmarking computer vision methods often provide accuracy rates above 90% in controlled settings. However, little focus has been given to aspects of fairness, uncertainty, and scalability within facial emotion recognition systems. The increasing applicability of FER models within assisted psychiatry and similar domains underlines the importance of fair and computational resource compliant decision-making. The primary objective of this paper is to propose methods for assessment of existing open source FER models to establish a thorough understanding of their current fairness, scalability, and robustness.

Cite this Paper

BibTeX
@InProceedings{pmlr-v233-fromberg24a, title = {Beyond Accuracy: Fairness, Scalability, and Uncertainty Considerations in Facial Emotion Recognition}, author = {Fromberg, Laurits and Nielsen, Troels and Frumosu, Flavia Dalia and Clemmensen, Line H}, booktitle = {Proceedings of the 5th Northern Lights Deep Learning Conference ({NLDL})}, pages = {67--74}, year = {2024}, editor = {Lutchyn, Tetiana and Ramírez Rivera, Adín and Ricaud, Benjamin}, volume = {233}, series = {Proceedings of Machine Learning Research}, month = {09--11 Jan}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v233/fromberg24a/fromberg24a.pdf}, url = {https://proceedings.mlr.press/v233/fromberg24a.html}, abstract = {Facial emotion recognition (FER) from images or videos is an emerging subfield of emotion recognition that in recent years has achieved increased traction resulting in a wide range of models, datasets, and applications. Benchmarking computer vision methods often provide accuracy rates above 90% in controlled settings. However, little focus has been given to aspects of fairness, uncertainty, and scalability within facial emotion recognition systems. The increasing applicability of FER models within assisted psychiatry and similar domains underlines the importance of fair and computational resource compliant decision-making. The primary objective of this paper is to propose methods for assessment of existing open source FER models to establish a thorough understanding of their current fairness, scalability, and robustness.} }
Endnote
%0 Conference Paper %T Beyond Accuracy: Fairness, Scalability, and Uncertainty Considerations in Facial Emotion Recognition %A Laurits Fromberg %A Troels Nielsen %A Flavia Dalia Frumosu %A Line H Clemmensen %B Proceedings of the 5th Northern Lights Deep Learning Conference ({NLDL}) %C Proceedings of Machine Learning Research %D 2024 %E Tetiana Lutchyn %E Adín Ramírez Rivera %E Benjamin Ricaud %F pmlr-v233-fromberg24a %I PMLR %P 67--74 %U https://proceedings.mlr.press/v233/fromberg24a.html %V 233 %X Facial emotion recognition (FER) from images or videos is an emerging subfield of emotion recognition that in recent years has achieved increased traction resulting in a wide range of models, datasets, and applications. Benchmarking computer vision methods often provide accuracy rates above 90% in controlled settings. However, little focus has been given to aspects of fairness, uncertainty, and scalability within facial emotion recognition systems. The increasing applicability of FER models within assisted psychiatry and similar domains underlines the importance of fair and computational resource compliant decision-making. The primary objective of this paper is to propose methods for assessment of existing open source FER models to establish a thorough understanding of their current fairness, scalability, and robustness.
APA
Fromberg, L., Nielsen, T., Frumosu, F.D. & Clemmensen, L.H.. (2024). Beyond Accuracy: Fairness, Scalability, and Uncertainty Considerations in Facial Emotion Recognition. Proceedings of the 5th Northern Lights Deep Learning Conference ({NLDL}), in Proceedings of Machine Learning Research 233:67-74 Available from https://proceedings.mlr.press/v233/fromberg24a.html.

Related Material