Babnik, Žiga ; Damer, Naser ; Štruc, Vitomir (2023)
Optimization-Based Improvement of Face Image Quality Assessment Techniques.
11th International Workshop on Biometrics and Forensics. Barcelona, Spain (19.-20.04.2023)
doi: 10.1109/IWBF57495.2023.10157796
Konferenzveröffentlichung, Bibliographie
Kurzbeschreibung (Abstract)
Contemporary face recognition (FR) models achieve near-ideal recognition performance in constrained settings, yet do not fully translate the performance to unconstrained (real-world) scenarios. To help improve the performance and stability of FR systems in such unconstrained settings, face image quality assessment (FIQA) techniques try to infer sample-quality information from the input face images that can aid with the recognition process. While existing FIQA techniques are able to efficiently capture the differences between high and low quality images, they typically cannot fully distinguish between images of similar quality, leading to lower performance in many scenarios. To address this issue, we present in this paper a supervised quality-label optimization approach, aimed at improving the performance of existing FIQA techniques. The developed optimization procedure infuses additional information (computed with a selected FR model) into the initial quality scores generated with a given FIQA technique to produce better estimates of the "actual" image quality. We evaluate the proposed approach in comprehensive experiments with six state-of-the-art FIQA approaches (CR-FIQA, FaceQAN, SER-FIQ, PCNet, MagFace, SER-FIQ) on five commonly used benchmarks (LFW, CFP-FP, CPLFW, CALFW, XQLFW) using three targeted FR models (ArcFace, ElasticFace, CurricularFace) with highly encouraging results.
| Typ des Eintrags: | Konferenzveröffentlichung |
|---|---|
| Erschienen: | 2023 |
| Autor(en): | Babnik, Žiga ; Damer, Naser ; Štruc, Vitomir |
| Art des Eintrags: | Bibliographie |
| Titel: | Optimization-Based Improvement of Face Image Quality Assessment Techniques |
| Sprache: | Englisch |
| Publikationsjahr: | 26 Juni 2023 |
| Verlag: | IEEE |
| Buchtitel: | 2023 11th International Workshop on Biometrics and Forensics (IWBF) |
| Veranstaltungstitel: | 11th International Workshop on Biometrics and Forensics |
| Veranstaltungsort: | Barcelona, Spain |
| Veranstaltungsdatum: | 19.-20.04.2023 |
| DOI: | 10.1109/IWBF57495.2023.10157796 |
| Kurzbeschreibung (Abstract): | Contemporary face recognition (FR) models achieve near-ideal recognition performance in constrained settings, yet do not fully translate the performance to unconstrained (real-world) scenarios. To help improve the performance and stability of FR systems in such unconstrained settings, face image quality assessment (FIQA) techniques try to infer sample-quality information from the input face images that can aid with the recognition process. While existing FIQA techniques are able to efficiently capture the differences between high and low quality images, they typically cannot fully distinguish between images of similar quality, leading to lower performance in many scenarios. To address this issue, we present in this paper a supervised quality-label optimization approach, aimed at improving the performance of existing FIQA techniques. The developed optimization procedure infuses additional information (computed with a selected FR model) into the initial quality scores generated with a given FIQA technique to produce better estimates of the "actual" image quality. We evaluate the proposed approach in comprehensive experiments with six state-of-the-art FIQA approaches (CR-FIQA, FaceQAN, SER-FIQ, PCNet, MagFace, SER-FIQ) on five commonly used benchmarks (LFW, CFP-FP, CPLFW, CALFW, XQLFW) using three targeted FR models (ArcFace, ElasticFace, CurricularFace) with highly encouraging results. |
| Freie Schlagworte: | Biometrics, Face recognition, Quality estimation, Machine learning, Deep learning |
| Fachbereich(e)/-gebiet(e): | 20 Fachbereich Informatik 20 Fachbereich Informatik > Graphisch-Interaktive Systeme |
| Hinterlegungsdatum: | 19 Jul 2023 07:18 |
| Letzte Änderung: | 19 Jul 2023 07:18 |
| PPN: | |
| Export: | |
| Suche nach Titel in: | TUfind oder in Google |
 |
Frage zum Eintrag |
Optionen (nur für Redakteure)
 |
Redaktionelle Details anzeigen |
Drucken
Impressum