On the Effectiveness of Image Rotation for Open Set Domain Adaptation

TL;DR

This paper introduces a novel open set domain adaptation method using image rotation recognition, which improves known class alignment and unknown rejection, validated by new metrics and benchmarks.

Contribution

It proposes a self-supervised rotation recognition approach for OSDA and introduces a new open set metric for fair evaluation.

Findings

Our method outperforms existing OSDA techniques.

Reproducibility is a key issue in OSDA research.

The new metric effectively balances known class recognition and unknown rejection.

Abstract

Open Set Domain Adaptation (OSDA) bridges the domain gap between a labeled source domain and an unlabeled target domain, while also rejecting target classes that are not present in the source. To avoid negative transfer, OSDA can be tackled by first separating the known/unknown target samples and then aligning known target samples with the source data. We propose a novel method to addresses both these problems using the self-supervised task of rotation recognition. Moreover, we assess the performance with a new open set metric that properly balances the contribution of recognizing the known classes and rejecting the unknown samples. Comparative experiments with existing OSDA methods on the standard Office-31 and Office-Home benchmarks show that: (i) our method outperforms its competitors, (ii) reproducibility for this field is a crucial issue to tackle, (iii) our metric provides a reliable tool to allow fair open set evaluation.