Unsupervised Face Synthesis Based on Human Traits

Roberto Leyva; Gregory Epiphaniou; Carsten Maple; Victor Sanchez

doi:10.1109/IWBF57495.2023.10157232

Unsupervised Face Synthesis Based on Human Traits

Roberto Leyva^*
, Gregory Epiphaniou
, Carsten Maple
, Victor Sanchez

^*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

3 Scopus citations

Abstract

This paper presents a strategy to synthesize face images based on human traits. Specifically, the strategy allows synthesizing face images with similar age, gender, and ethnicity, after discovering groups of people with similar facial features. Our synthesizer is based on unsupervised learning and is capable to generate realistic faces. Our experiments reveal that grouping the training samples according to their similarity can lead to more realistic face images while having semantic control over the synthesis. The proposed strategy achieves competitive performance compared to the state-of-the-art and outperforms the baseline in terms of the Frechet Inception Distance.

Original language	English
Title of host publication	2023 11th International Workshop on Biometrics and Forensics, IWBF 2023
Publisher	Institute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)	9798350336078
DOIs	https://doi.org/10.1109/IWBF57495.2023.10157232
State	Published - 2023
Externally published	Yes
Event	11th International Workshop on Biometrics and Forensics, IWBF 2023 - Barcelona, Spain Duration: 19 Apr 2023 → 20 Apr 2023

Publication series

Name	2023 11th International Workshop on Biometrics and Forensics, IWBF 2023

Conference

Conference	11th International Workshop on Biometrics and Forensics, IWBF 2023
Country/Territory	Spain
City	Barcelona
Period	19/04/23 → 20/04/23

Bibliographical note

Publisher Copyright:
© 2023 IEEE.

Keywords

biometrics
Face synthesis
mixture models
unsupervised learning

ASJC Scopus subject areas

Computer Vision and Pattern Recognition
Pathology and Forensic Medicine
Biomedical Engineering
Instrumentation

Access to Document

10.1109/IWBF57495.2023.10157232

Cite this

@inproceedings{683867239e31409d928472268e1a4c17,

title = "Unsupervised Face Synthesis Based on Human Traits",

abstract = "This paper presents a strategy to synthesize face images based on human traits. Specifically, the strategy allows synthesizing face images with similar age, gender, and ethnicity, after discovering groups of people with similar facial features. Our synthesizer is based on unsupervised learning and is capable to generate realistic faces. Our experiments reveal that grouping the training samples according to their similarity can lead to more realistic face images while having semantic control over the synthesis. The proposed strategy achieves competitive performance compared to the state-of-the-art and outperforms the baseline in terms of the Frechet Inception Distance.",

keywords = "biometrics, Face synthesis, mixture models, unsupervised learning",

author = "Roberto Leyva and Gregory Epiphaniou and Carsten Maple and Victor Sanchez",

note = "Publisher Copyright: {\textcopyright} 2023 IEEE.; 11th International Workshop on Biometrics and Forensics, IWBF 2023 ; Conference date: 19-04-2023 Through 20-04-2023",

year = "2023",

doi = "10.1109/IWBF57495.2023.10157232",

language = "English",

series = "2023 11th International Workshop on Biometrics and Forensics, IWBF 2023",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

booktitle = "2023 11th International Workshop on Biometrics and Forensics, IWBF 2023",

address = "United States",

}

Leyva, R, Epiphaniou, G, Maple, C & Sanchez, V 2023, Unsupervised Face Synthesis Based on Human Traits. in 2023 11th International Workshop on Biometrics and Forensics, IWBF 2023. 2023 11th International Workshop on Biometrics and Forensics, IWBF 2023, Institute of Electrical and Electronics Engineers Inc., 11th International Workshop on Biometrics and Forensics, IWBF 2023, Barcelona, Spain, 19/04/23. https://doi.org/10.1109/IWBF57495.2023.10157232

Unsupervised Face Synthesis Based on Human Traits. / Leyva, Roberto; Epiphaniou, Gregory; Maple, Carsten et al.
2023 11th International Workshop on Biometrics and Forensics, IWBF 2023. Institute of Electrical and Electronics Engineers Inc., 2023. (2023 11th International Workshop on Biometrics and Forensics, IWBF 2023).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Unsupervised Face Synthesis Based on Human Traits

AU - Leyva, Roberto

AU - Epiphaniou, Gregory

AU - Maple, Carsten

AU - Sanchez, Victor

PY - 2023

Y1 - 2023

N2 - This paper presents a strategy to synthesize face images based on human traits. Specifically, the strategy allows synthesizing face images with similar age, gender, and ethnicity, after discovering groups of people with similar facial features. Our synthesizer is based on unsupervised learning and is capable to generate realistic faces. Our experiments reveal that grouping the training samples according to their similarity can lead to more realistic face images while having semantic control over the synthesis. The proposed strategy achieves competitive performance compared to the state-of-the-art and outperforms the baseline in terms of the Frechet Inception Distance.

AB - This paper presents a strategy to synthesize face images based on human traits. Specifically, the strategy allows synthesizing face images with similar age, gender, and ethnicity, after discovering groups of people with similar facial features. Our synthesizer is based on unsupervised learning and is capable to generate realistic faces. Our experiments reveal that grouping the training samples according to their similarity can lead to more realistic face images while having semantic control over the synthesis. The proposed strategy achieves competitive performance compared to the state-of-the-art and outperforms the baseline in terms of the Frechet Inception Distance.

KW - biometrics

KW - Face synthesis

KW - mixture models

KW - unsupervised learning

UR - https://www.scopus.com/pages/publications/85165687960

U2 - 10.1109/IWBF57495.2023.10157232

DO - 10.1109/IWBF57495.2023.10157232

M3 - Conference contribution

AN - SCOPUS:85165687960

T3 - 2023 11th International Workshop on Biometrics and Forensics, IWBF 2023

BT - 2023 11th International Workshop on Biometrics and Forensics, IWBF 2023

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 11th International Workshop on Biometrics and Forensics, IWBF 2023

Y2 - 19 April 2023 through 20 April 2023

ER -

Unsupervised Face Synthesis Based on Human Traits

Abstract

Publication series

Conference

Bibliographical note

Keywords

ASJC Scopus subject areas

Access to Document

Fingerprint

Cite this