Loading [MathJax]/extensions/MathMenu.js
Reducing Head Pose Estimation Data Set Bias With Synthetic Data | IEEE Journals & Magazine | IEEE Xplore

Reducing Head Pose Estimation Data Set Bias With Synthetic Data


In this work, we advocate for the use of a controllable generative model to create synthetic images for HPE, offering an efficient and cost-effective alternative, improvi...

Abstract:

Data set bias not only compromises the fairness, accuracy and effectiveness of trained models, but also leads to a lower performance in real-world scenarios compared to t...Show More

Abstract:

Data set bias not only compromises the fairness, accuracy and effectiveness of trained models, but also leads to a lower performance in real-world scenarios compared to the evaluation results obtained with a specific data set. This issue is especially evident in the estimation of head pose, as current data sets suffer from a limited number of images, imbalanced data distributions, the high cost of annotation, and ethical concerns. Synthetic data offers a promising solution to address these challenges, but current semi-synthetic data sets fail to deliver satisfactory results, likely due to the limited realism of the generated faces and the heavily skewed pose distribution. In this paper, we report the existence of data set biases in the most widely used head pose estimation benchmarks, which lead to an optimistic estimation of model performance in real-world scenarios. To mitigate this issue, we create a synthetic image data set using a generative model with explicit control over the head pose. Our experiments demonstrate that incorporating our synthetic images leads to improved generalization and accuracy.
In this work, we advocate for the use of a controllable generative model to create synthetic images for HPE, offering an efficient and cost-effective alternative, improvi...
Published in: IEEE Access ( Volume: 13)
Page(s): 73530 - 73539
Date of Publication: 16 April 2025
Electronic ISSN: 2169-3536

Funding Agency:

Author image of Roberto Valle
Departamento de Inteligencia Artificial, Universidad Politécnica de Madrid, Madrid, Spain
Roberto Valle received the integrated B.S. and M.S. degrees from the Universidad Rey Juan Carlos (URJC), in 2013, and the Ph.D. degree in artificial intelligence from the Universidad Politécnica de Madrid (UPM), in 2020. Since 2022, he has been an Assistant Professor with UPM and a member of the Computer Perception Group. His research interests include computer vision and deep learning algorithms for facial analysis probl...Show More
Roberto Valle received the integrated B.S. and M.S. degrees from the Universidad Rey Juan Carlos (URJC), in 2013, and the Ph.D. degree in artificial intelligence from the Universidad Politécnica de Madrid (UPM), in 2020. Since 2022, he has been an Assistant Professor with UPM and a member of the Computer Perception Group. His research interests include computer vision and deep learning algorithms for facial analysis probl...View more
Author image of José M. Buenaposada
Departamento de Informática y Estadística, Universidad Rey Juan Carlos, Móstoles, Madrid, Spain
José M. Buenaposada received the integrated B.S. and M.S. and the Ph.D. degrees in computer science from the Universidad Politécnica de Madrid (UPM), in 1999 and 2005, respectively. From 2008 to 2018, he was an Associate Professor with the Universidad Rey Juan Carlos (URJC), Spain, where he has been a Professor of computer science, since July 2023. He is a member of the GAVAB Research Group, URJC, and the Computer Percept...Show More
José M. Buenaposada received the integrated B.S. and M.S. and the Ph.D. degrees in computer science from the Universidad Politécnica de Madrid (UPM), in 1999 and 2005, respectively. From 2008 to 2018, he was an Associate Professor with the Universidad Rey Juan Carlos (URJC), Spain, where he has been a Professor of computer science, since July 2023. He is a member of the GAVAB Research Group, URJC, and the Computer Percept...View more
Author image of Luis Baumela
Departamento de Inteligencia Artificial, Universidad Politécnica de Madrid, Madrid, Spain
Luis Baumela received the integrated B.S. and M.S. and the Ph.D. degrees in computer science from the Universidad Politécnica de Madrid, in 1989 and 1995, respectively. From 1989 to 1992, he was an Engineer with Telefónica’s Research and Development Laboratories. From 1997 to 2016, he was an Associate Professor. Since 2016, he has been a Professor of computer science with the ETSI Informáticos, Universidad Politécnica de ...Show More
Luis Baumela received the integrated B.S. and M.S. and the Ph.D. degrees in computer science from the Universidad Politécnica de Madrid, in 1989 and 1995, respectively. From 1989 to 1992, he was an Engineer with Telefónica’s Research and Development Laboratories. From 1997 to 2016, he was an Associate Professor. Since 2016, he has been a Professor of computer science with the ETSI Informáticos, Universidad Politécnica de ...View more

Author image of Roberto Valle
Departamento de Inteligencia Artificial, Universidad Politécnica de Madrid, Madrid, Spain
Roberto Valle received the integrated B.S. and M.S. degrees from the Universidad Rey Juan Carlos (URJC), in 2013, and the Ph.D. degree in artificial intelligence from the Universidad Politécnica de Madrid (UPM), in 2020. Since 2022, he has been an Assistant Professor with UPM and a member of the Computer Perception Group. His research interests include computer vision and deep learning algorithms for facial analysis problems.
Roberto Valle received the integrated B.S. and M.S. degrees from the Universidad Rey Juan Carlos (URJC), in 2013, and the Ph.D. degree in artificial intelligence from the Universidad Politécnica de Madrid (UPM), in 2020. Since 2022, he has been an Assistant Professor with UPM and a member of the Computer Perception Group. His research interests include computer vision and deep learning algorithms for facial analysis problems.View more
Author image of José M. Buenaposada
Departamento de Informática y Estadística, Universidad Rey Juan Carlos, Móstoles, Madrid, Spain
José M. Buenaposada received the integrated B.S. and M.S. and the Ph.D. degrees in computer science from the Universidad Politécnica de Madrid (UPM), in 1999 and 2005, respectively. From 2008 to 2018, he was an Associate Professor with the Universidad Rey Juan Carlos (URJC), Spain, where he has been a Professor of computer science, since July 2023. He is a member of the GAVAB Research Group, URJC, and the Computer Perception Group, UPM. His research interests include image alignment, face image analysis, object detection, and efficient computer vision.
José M. Buenaposada received the integrated B.S. and M.S. and the Ph.D. degrees in computer science from the Universidad Politécnica de Madrid (UPM), in 1999 and 2005, respectively. From 2008 to 2018, he was an Associate Professor with the Universidad Rey Juan Carlos (URJC), Spain, where he has been a Professor of computer science, since July 2023. He is a member of the GAVAB Research Group, URJC, and the Computer Perception Group, UPM. His research interests include image alignment, face image analysis, object detection, and efficient computer vision.View more
Author image of Luis Baumela
Departamento de Inteligencia Artificial, Universidad Politécnica de Madrid, Madrid, Spain
Luis Baumela received the integrated B.S. and M.S. and the Ph.D. degrees in computer science from the Universidad Politécnica de Madrid, in 1989 and 1995, respectively. From 1989 to 1992, he was an Engineer with Telefónica’s Research and Development Laboratories. From 1997 to 2016, he was an Associate Professor. Since 2016, he has been a Professor of computer science with the ETSI Informáticos, Universidad Politécnica de Madrid, where he leads the Computer Perception Group. His research interest includes the development of ethical vision systems that look at people.
Luis Baumela received the integrated B.S. and M.S. and the Ph.D. degrees in computer science from the Universidad Politécnica de Madrid, in 1989 and 1995, respectively. From 1989 to 1992, he was an Engineer with Telefónica’s Research and Development Laboratories. From 1997 to 2016, he was an Associate Professor. Since 2016, he has been a Professor of computer science with the ETSI Informáticos, Universidad Politécnica de Madrid, where he leads the Computer Perception Group. His research interest includes the development of ethical vision systems that look at people.View more

References

References is not available for this document.