Effectiveness of Blind Face Restoration to Boost Face Recognition Performance at Low-Resolution Images
Published in International Workshop on Artificial Intelligence and Pattern Recognition, 2023
Recommended citation: Yoanna Martínez-Díaz, Luis S. Luevano, Heydi Méndez-Vázquez. "Effectiveness of Blind Face Restoration to Boost Face Recognition Performance at Low-Resolution Images". Progress in Artificial Intelligence and Pattern Recognition. IWAIPR 2023. Lecture Notes in Computer Science, vol 14335. Springer, Cham. https://doi.org/10.1007/978-3-031-49552-6_39 https://link.springer.com/chapter/10.1007/978-3-031-49552-6_39
This paper studies the effectiveness of Blind Face Restoration methods to boost the performance of face recognition systems on low-resolution images. We investigate the use of three blind face restoration techniques, which have demonstrated impressive results in generating realistic high-resolution face images. Three state-of-the-art face recognition methods were selected to assess the impact of using the generated high-resolution images on their performance. Our analysis includes both, synthesized and native low-resolution images. The conducted experimental evaluation show that this is still an open research problem.
Latex citation:
@InProceedings{10.1007/978-3-031-49552-6_39,
author="Mart{\'i}nez-D{\'i}az, Yoanna
and Lu{\'e}vano, Luis S.
and M{\'e}ndez-V{\'a}zquez, Heydi",
editor="Hern{\'a}ndez Heredia, Yanio
and Mili{\'a}n N{\'u}{\~{n}}ez, Vladimir
and Ruiz Shulcloper, Jos{\'e}",
title="Effectiveness of Blind Face Restoration to Boost Face Recognition Performance at Low-Resolution Images",
booktitle="Progress in Artificial Intelligence and Pattern Recognition",
year="2024",
publisher="Springer Nature Switzerland",
address="Cham",
pages="455--467",
abstract="This paper studies the effectiveness of Blind Face Restoration methods to boost the performance of face recognition systems on low-resolution images. We investigate the use of three blind face restoration techniques, which have demonstrated impressive results in generating realistic high-resolution face images. Three state-of-the-art face recognition methods were selected to assess the impact of using the generated high-resolution images on their performance. Our analysis includes both, synthesized and native low-resolution images. The conducted experimental evaluation show that this is still an open research problem.",
isbn="978-3-031-49552-6"
}