Catastrophic Interference in Disguised Face Recognition

Parichehr B. Ardakani*, Diego Velazquez, Josep M. Gonfaus, Pau Rodríguez, F. Xavier Roca, Jordi Gonzàlez

*Corresponding author for this work

Research output: Chapter in BookChapterResearchpeer-review

Abstract

It is commonly known the natural tendency of artificial neural networks to completely and abruptly forget previously known information when learning new information. We explore this behaviour in the context of Face Verification on the recently proposed Disguised Faces in the Wild dataset (DFW). We empirically evaluate several commonly used DCNN architectures on Face Recognition and distill some insights about the effect of sequential learning on distinct identities from different datasets, showing that the catastrophic forgetness phenomenon is present even in feature embeddings fine-tuned on different tasks from the original domain.

Original languageEnglish
Title of host publicationPattern Recognition and Image Analysis - 9th Iberian Conference, IbPRIA 2019, Proceedings
EditorsAythami Morales, Julian Fierrez, José Salvador Sánchez, Bernardete Ribeiro
Pages64-75
Number of pages12
DOIs
Publication statusPublished - 2019

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume11868 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Keywords

  • Disguised Faces
  • Face recognition
  • Neural network forgetness

Fingerprint

Dive into the research topics of 'Catastrophic Interference in Disguised Face Recognition'. Together they form a unique fingerprint.

Cite this