Recent Advances in Transfer Learning for Cross-Dataset Visual Recognition: A Problem-Oriented Perspective

TL;DR

This paper provides a comprehensive, problem-oriented review of transfer learning methods for cross-dataset visual recognition, categorizing 17 problems and analyzing research progress and gaps.

Contribution

It introduces a novel problem-oriented taxonomy for cross-dataset recognition and systematically reviews transfer learning approaches for each problem.

Findings

Identifies 17 distinct cross-dataset recognition problems.

Highlights that 8 problems are scarcely studied.

Provides a systematic categorization and analysis of transfer learning methods.

Abstract

This paper takes a problem-oriented perspective and presents a comprehensive review of transfer learning methods, both shallow and deep, for cross-dataset visual recognition. Specifically, it categorises the cross-dataset recognition into seventeen problems based on a set of carefully chosen data and label attributes. Such a problem-oriented taxonomy has allowed us to examine how different transfer learning approaches tackle each problem and how well each problem has been researched to date. The comprehensive problem-oriented review of the advances in transfer learning with respect to the problem has not only revealed the challenges in transfer learning for visual recognition, but also the problems (e.g. eight of the seventeen problems) that have been scarcely studied. This survey not only presents an up-to-date technical review for researchers, but also a systematic approach and a reference for a machine learning practitioner to categorise a real problem and to look up for a possible solution accordingly.