Face alignment: improving the accuracy of fast models using domain-specific unlabelled data and a teacher–student scheme

Constantino Alvarez Casado, Miguel Bordallo Lopez

Research output: Contribution to journalArticleScientificpeer-review

Abstract

Face alignment is a crucial step in multiple face analysis and recognition tasks. The current state-of-the-art is comprised by very slow methods based on deep learning that require computationally heavy inference and very fast methods based on cascades of regressors that lack the ability to cope with complicated cases or extreme poses. The authors show how collecting a small subset of unlabelled domain-specific data can improve the accuracy of fast-inference models utilising data annotated by a slower one and a teacher–student architecture. In the proposed solution, they annotate a small subset of facial images belonging to two challenging domains using a slow but more accurate model, and this data is used to incrementally train a fast one. Their results show that by adding as little as a 5% of challenging data, they can reduce the error rate in a specific domain up to 30% without losing any generalisation abilities. This training scheme has applicability in numerous computer vision and engineering problems where computational power and model size are constrained by the application and platform or real-time operation is a requirement.
Original languageEnglish
Pages (from-to)646 - 648
Number of pages3
JournalElectronics Letters
Volume55
Issue number11
DOIs
Publication statusPublished - 2019
MoE publication typeA1 Journal article-refereed

Fingerprint

Computer vision
Deep learning

Keywords

  • computer vision
  • learning (artificial intelligence)
  • face recognition
  • engineering problems
  • computational power
  • model size
  • numerous computer vision
  • training scheme
  • generalisation abilities
  • specific domain
  • challenging data
  • slow but more accurate model
  • challenging domains
  • teacher–student architecture
  • fast-inference models
  • unlabelled domain-specific data
  • extreme poses
  • complicated cases
  • fast methods
  • computationally heavy inference
  • deep learning
  • slow methods
  • current state-of-the-art
  • recognition tasks
  • multiple face analysis
  • crucial step
  • teacher–student scheme
  • domain-specific unlabelled data
  • fast models
  • face alignment

Cite this

@article{8a02e6e6d348409ab0cc68de954da658,
title = "Face alignment: improving the accuracy of fast models using domain-specific unlabelled data and a teacher–student scheme",
abstract = "Face alignment is a crucial step in multiple face analysis and recognition tasks. The current state-of-the-art is comprised by very slow methods based on deep learning that require computationally heavy inference and very fast methods based on cascades of regressors that lack the ability to cope with complicated cases or extreme poses. The authors show how collecting a small subset of unlabelled domain-specific data can improve the accuracy of fast-inference models utilising data annotated by a slower one and a teacher–student architecture. In the proposed solution, they annotate a small subset of facial images belonging to two challenging domains using a slow but more accurate model, and this data is used to incrementally train a fast one. Their results show that by adding as little as a 5{\%} of challenging data, they can reduce the error rate in a specific domain up to 30{\%} without losing any generalisation abilities. This training scheme has applicability in numerous computer vision and engineering problems where computational power and model size are constrained by the application and platform or real-time operation is a requirement.",
keywords = "computer vision, learning (artificial intelligence), face recognition, engineering problems, computational power, model size, numerous computer vision, training scheme, generalisation abilities, specific domain, challenging data, slow but more accurate model, challenging domains, teacher–student architecture, fast-inference models, unlabelled domain-specific data, extreme poses, complicated cases, fast methods, computationally heavy inference, deep learning, slow methods, current state-of-the-art, recognition tasks, multiple face analysis, crucial step, teacher–student scheme, domain-specific unlabelled data, fast models, face alignment",
author = "{Alvarez Casado}, Constantino and {Bordallo Lopez}, Miguel",
year = "2019",
doi = "10.1049/el.2019.0319",
language = "English",
volume = "55",
pages = "646 -- 648",
journal = "Electronics Letters",
issn = "0013-5194",
publisher = "Institution of Engineering and Technology IET",
number = "11",

}

Face alignment : improving the accuracy of fast models using domain-specific unlabelled data and a teacher–student scheme. / Alvarez Casado, Constantino; Bordallo Lopez, Miguel.

In: Electronics Letters, Vol. 55, No. 11, 2019, p. 646 - 648.

Research output: Contribution to journalArticleScientificpeer-review

TY - JOUR

T1 - Face alignment

T2 - improving the accuracy of fast models using domain-specific unlabelled data and a teacher–student scheme

AU - Alvarez Casado, Constantino

AU - Bordallo Lopez, Miguel

PY - 2019

Y1 - 2019

N2 - Face alignment is a crucial step in multiple face analysis and recognition tasks. The current state-of-the-art is comprised by very slow methods based on deep learning that require computationally heavy inference and very fast methods based on cascades of regressors that lack the ability to cope with complicated cases or extreme poses. The authors show how collecting a small subset of unlabelled domain-specific data can improve the accuracy of fast-inference models utilising data annotated by a slower one and a teacher–student architecture. In the proposed solution, they annotate a small subset of facial images belonging to two challenging domains using a slow but more accurate model, and this data is used to incrementally train a fast one. Their results show that by adding as little as a 5% of challenging data, they can reduce the error rate in a specific domain up to 30% without losing any generalisation abilities. This training scheme has applicability in numerous computer vision and engineering problems where computational power and model size are constrained by the application and platform or real-time operation is a requirement.

AB - Face alignment is a crucial step in multiple face analysis and recognition tasks. The current state-of-the-art is comprised by very slow methods based on deep learning that require computationally heavy inference and very fast methods based on cascades of regressors that lack the ability to cope with complicated cases or extreme poses. The authors show how collecting a small subset of unlabelled domain-specific data can improve the accuracy of fast-inference models utilising data annotated by a slower one and a teacher–student architecture. In the proposed solution, they annotate a small subset of facial images belonging to two challenging domains using a slow but more accurate model, and this data is used to incrementally train a fast one. Their results show that by adding as little as a 5% of challenging data, they can reduce the error rate in a specific domain up to 30% without losing any generalisation abilities. This training scheme has applicability in numerous computer vision and engineering problems where computational power and model size are constrained by the application and platform or real-time operation is a requirement.

KW - computer vision

KW - learning (artificial intelligence)

KW - face recognition

KW - engineering problems

KW - computational power

KW - model size

KW - numerous computer vision

KW - training scheme

KW - generalisation abilities

KW - specific domain

KW - challenging data

KW - slow but more accurate model

KW - challenging domains

KW - teacher–student architecture

KW - fast-inference models

KW - unlabelled domain-specific data

KW - extreme poses

KW - complicated cases

KW - fast methods

KW - computationally heavy inference

KW - deep learning

KW - slow methods

KW - current state-of-the-art

KW - recognition tasks

KW - multiple face analysis

KW - crucial step

KW - teacher–student scheme

KW - domain-specific unlabelled data

KW - fast models

KW - face alignment

UR - http://www.scopus.com/inward/record.url?scp=85066733529&partnerID=8YFLogxK

U2 - 10.1049/el.2019.0319

DO - 10.1049/el.2019.0319

M3 - Article

VL - 55

SP - 646

EP - 648

JO - Electronics Letters

JF - Electronics Letters

SN - 0013-5194

IS - 11

ER -