On the Effectiveness of Image Rotation for Open Set Domain Adaptation
This work addresses domain adaptation in scenarios where target domains contain unknown classes, a problem for computer vision applications, though it is incremental as it builds on existing OSDA frameworks.
The paper tackles open set domain adaptation by using image rotation as a self-supervised task to separate known and unknown target samples and align known ones with the source, achieving superior performance on Office-31 and Office-Home benchmarks compared to existing methods.
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.