Tampere University of Technology

TUTCRIS Research Portal

Multi-task Deep Face Recognition

Research output: Chapter in Book/Report/Conference proceedingConference contributionScientificpeer-review

Details

Original languageEnglish
Title of host publicationBiometric Recognition - 12th Chinese Conference, CCBR 2017, Proceedings
PublisherSpringer Verlag
Pages183-190
Number of pages8
ISBN (Print)9783319699226
DOIs
Publication statusPublished - 2017
Publication typeA4 Article in a conference publication
EventChinese Conference on Biometric Recognition -
Duration: 1 Jan 2000 → …

Publication series

NameLecture Notes in Computer Science
Volume10568
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

ConferenceChinese Conference on Biometric Recognition
Period1/01/00 → …

Abstract

In recent years, deep learning has become one of the most representative and effective techniques in face recognition. Due to the high expense of labelling data, it is costly to collect a large-scale face dataset with accurate label information. For the tasks without sufficient data, deep models cannot be well trained. Generally, parameters of deep models are usually initialized with a pre-trained model, and then fine-tuned on a small dataset of specific task. However, by straightforward fine-tuning, the final model usually does not generalize well. In this paper, we propose a multi-task deep learning (MTDL) method for face recognition. The superiority of the proposed multi-task method is demonstrated by experiments on LFW and CCFD.

Keywords

  • Convolution neural network, Deep learning, Face recognition, Multi-task

Publication forum classification

Field of science, Statistics Finland