Masked-to-Unmasked Face Translation: A Review of Datasets, Approaches, and Metrics

Authors

  • Yazid Safiny IDIA Laboratory, National School of Applied Sciences, Sultan Moulay Slimane University, Khouribga, Morocco
  • Siham Ahmam IDIA Laboratory, National School of Applied Sciences Khouribga, Sultan Moulay Slimane University, Morocco
  • Nidal Lamghari IDIA Laboratory, National School of Applied Sciences Khouribga, Sultan Moulay Slimane University, Morocco
  • Abdelghani Ghazdali LaMDa Laboratory, National School of Applied Sciences Khouribga, Sultan Moulay Slimane University, Morocco

DOI:

https://doi.org/10.19139/soic-2310-5070-4153

Keywords:

Masked Face Recognition, Face Mask Removal, Face Unmasking, Face de-occlusion

Abstract

The emergence of face masks is causing major issues for computer vision systems as they prevent access to crucial parts of faces needed for applications like face recognition and expression analysis. To solve this issue, face unmasking which is a process that involves removing face masks and generating hidden facial details has become a popular topic of investigation. This paper provides an overview of various solutions proposed for face unmasking, paying particular attention to deep generative models, mainly based on GANs. Face unmasking methods can be grouped into three categories: one-stage networks, two-stage networks, and image inpainting techniques. Advantages, limitations, and architectural characteristics of each category will be analyzed comparatively. In addition, this paper looks at frequently utilized data sources and analyzes problems caused by synthetic mask generation. Metrics for evaluating face unmasking systems and losses used in them will be assessed in relation to realism and consistency of generated faces. Moreover, issues regarding benchmarking, reproducibility, and ethics of this area of research will be addressed. Future directions of the field will be outlined as well.

Downloads

Published

2026-10-02

How to Cite

Safiny, Y., Ahmam, S., Lamghari, N., & Ghazdali, A. (2026). Masked-to-Unmasked Face Translation: A Review of Datasets, Approaches, and Metrics. Statistics, Optimization & Information Computing. https://doi.org/10.19139/soic-2310-5070-4153

Issue

Online First

Section

ICMICSA-V2 2025

Categories

License

Copyright (c) 2026 Yazid Safiny, Siham Ahmam, Nidal Lamghari, Abdelghani Ghazdali

Creative Commons License

This work is licensed under a Creative Commons Attribution 4.0 International License.

Authors who publish with this journal agree to the following terms:

    1. Authors retain copyright and grant the journal right of first publication with the work simultaneously licensed under a Creative Commons Attribution License that allows others to share the work with an acknowledgement of the work's authorship and initial publication in this journal.
    2. Authors are able to enter into separate, additional contractual arrangements for the non-exclusive distribution of the journal's published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgement of its initial publication in this journal.
    3. Authors are permitted and encouraged to post their work online (e.g., in institutional repositories or on their website) prior to and during the submission process, as it can lead to productive exchanges, as well as earlier and greater citation of published work (See The Effect of Open Access).