Machine Learning Approaches for Prediction of Facial Rejuvenation Using Real and Synthetic Data

Syed Afaq Ali Shah, Mohammed Bennamoun, Michael K. Molton

Research output: Contribution to journalArticle

Abstract

This paper proposes a novel machine learning approaches to predict the outcome of facial rejuvenation prior to a cosmetic procedure. This is achieved by estimating the required amount of dermal filler volume that needs to be applied on the face by learning the underlying structural mapping from the pretreatment and posttreatment 3D face images. We develop and train our proposed deep neural network, called Rejuv3DNet, designed specifically to predict the dermal filler volume. We also propose the kernel regression (KR)-based model to validate and improve our volume estimation results using regression. Our other contributions include the development of the first 3D face cosmetic dataset, which consists of real-world pretreatment and posttreatment 3D face images and a novel technique for the generation of synthetic cosmetic treatment 3D face images. Our experimental results show that the proposed Rejuv3DNet and the KR model achieve 62.5% and 66.67%, respectively, on real-world data, while these techniques achieve a prediction accuracy of 75.2% and 89.5%, and 77.2% and 90.1% on our two different synthetic datasets. Our proposed techniques have been found to be computationally efficient, achieving near real-time prediction performance. The reported accuracies are our preliminary results for proof of concept, which can be improved with more data. The proposed approach has the potential for further investigation in the cosmetic surgery domain.

Original languageEnglish
Pages (from-to)23779-23787
Number of pages9
JournalIEEE Access
Volume7
DOIs
Publication statusPublished - 2019

Cite this

@article{6b33a42e88d54cb0a3db3e74ff77727a,
title = "Machine Learning Approaches for Prediction of Facial Rejuvenation Using Real and Synthetic Data",
abstract = "This paper proposes a novel machine learning approaches to predict the outcome of facial rejuvenation prior to a cosmetic procedure. This is achieved by estimating the required amount of dermal filler volume that needs to be applied on the face by learning the underlying structural mapping from the pretreatment and posttreatment 3D face images. We develop and train our proposed deep neural network, called Rejuv3DNet, designed specifically to predict the dermal filler volume. We also propose the kernel regression (KR)-based model to validate and improve our volume estimation results using regression. Our other contributions include the development of the first 3D face cosmetic dataset, which consists of real-world pretreatment and posttreatment 3D face images and a novel technique for the generation of synthetic cosmetic treatment 3D face images. Our experimental results show that the proposed Rejuv3DNet and the KR model achieve 62.5{\%} and 66.67{\%}, respectively, on real-world data, while these techniques achieve a prediction accuracy of 75.2{\%} and 89.5{\%}, and 77.2{\%} and 90.1{\%} on our two different synthetic datasets. Our proposed techniques have been found to be computationally efficient, achieving near real-time prediction performance. The reported accuracies are our preliminary results for proof of concept, which can be improved with more data. The proposed approach has the potential for further investigation in the cosmetic surgery domain.",
keywords = "Deep learning, deep neural network, facial analysis, regression, FACE RECOGNITION, EXPRESSION RECOGNITION, FEATURES, WRINKLE, SCALE, MODEL",
author = "Shah, {Syed Afaq Ali} and Mohammed Bennamoun and Molton, {Michael K.}",
year = "2019",
doi = "10.1109/ACCESS.2019.2899379",
language = "English",
volume = "7",
pages = "23779--23787",
journal = "IEEE Access",
issn = "2169-3536",
publisher = "IEEE, Institute of Electrical and Electronics Engineers",

}

Machine Learning Approaches for Prediction of Facial Rejuvenation Using Real and Synthetic Data. / Shah, Syed Afaq Ali; Bennamoun, Mohammed; Molton, Michael K.

In: IEEE Access, Vol. 7, 2019, p. 23779-23787.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Machine Learning Approaches for Prediction of Facial Rejuvenation Using Real and Synthetic Data

AU - Shah, Syed Afaq Ali

AU - Bennamoun, Mohammed

AU - Molton, Michael K.

PY - 2019

Y1 - 2019

N2 - This paper proposes a novel machine learning approaches to predict the outcome of facial rejuvenation prior to a cosmetic procedure. This is achieved by estimating the required amount of dermal filler volume that needs to be applied on the face by learning the underlying structural mapping from the pretreatment and posttreatment 3D face images. We develop and train our proposed deep neural network, called Rejuv3DNet, designed specifically to predict the dermal filler volume. We also propose the kernel regression (KR)-based model to validate and improve our volume estimation results using regression. Our other contributions include the development of the first 3D face cosmetic dataset, which consists of real-world pretreatment and posttreatment 3D face images and a novel technique for the generation of synthetic cosmetic treatment 3D face images. Our experimental results show that the proposed Rejuv3DNet and the KR model achieve 62.5% and 66.67%, respectively, on real-world data, while these techniques achieve a prediction accuracy of 75.2% and 89.5%, and 77.2% and 90.1% on our two different synthetic datasets. Our proposed techniques have been found to be computationally efficient, achieving near real-time prediction performance. The reported accuracies are our preliminary results for proof of concept, which can be improved with more data. The proposed approach has the potential for further investigation in the cosmetic surgery domain.

AB - This paper proposes a novel machine learning approaches to predict the outcome of facial rejuvenation prior to a cosmetic procedure. This is achieved by estimating the required amount of dermal filler volume that needs to be applied on the face by learning the underlying structural mapping from the pretreatment and posttreatment 3D face images. We develop and train our proposed deep neural network, called Rejuv3DNet, designed specifically to predict the dermal filler volume. We also propose the kernel regression (KR)-based model to validate and improve our volume estimation results using regression. Our other contributions include the development of the first 3D face cosmetic dataset, which consists of real-world pretreatment and posttreatment 3D face images and a novel technique for the generation of synthetic cosmetic treatment 3D face images. Our experimental results show that the proposed Rejuv3DNet and the KR model achieve 62.5% and 66.67%, respectively, on real-world data, while these techniques achieve a prediction accuracy of 75.2% and 89.5%, and 77.2% and 90.1% on our two different synthetic datasets. Our proposed techniques have been found to be computationally efficient, achieving near real-time prediction performance. The reported accuracies are our preliminary results for proof of concept, which can be improved with more data. The proposed approach has the potential for further investigation in the cosmetic surgery domain.

KW - Deep learning

KW - deep neural network

KW - facial analysis

KW - regression

KW - FACE RECOGNITION

KW - EXPRESSION RECOGNITION

KW - FEATURES

KW - WRINKLE

KW - SCALE

KW - MODEL

U2 - 10.1109/ACCESS.2019.2899379

DO - 10.1109/ACCESS.2019.2899379

M3 - Article

VL - 7

SP - 23779

EP - 23787

JO - IEEE Access

JF - IEEE Access

SN - 2169-3536

ER -