Multi-layer Domain Adaptation for Deep Convolutional Networks
This addresses the challenge of adopting deep learning in clinical settings with scarce and variable imaging data, representing an incremental improvement over existing domain adaptation techniques.
The paper tackles the problem of deep convolutional networks requiring large labeled datasets and failing on unseen domains, by proposing a domain adaptation method using gradient reversal layers and Squeeze-and-Excite modules, which improved multi-class classification accuracy by 5-20% compared to DANN on histopathology and chest X-ray databases.
Despite their success in many computer vision tasks, convolutional networks tend to require large amounts of labeled data to achieve generalization. Furthermore, the performance is not guaranteed on a sample from an unseen domain at test time, if the network was not exposed to similar samples from that domain at training time. This hinders the adoption of these techniques in clinical setting where the imaging data is scarce, and where the intra- and inter-domain variance of the data can be substantial. We propose a domain adaptation technique that is especially suitable for deep networks to alleviate this requirement of labeled data. Our method utilizes gradient reversal layers and Squeezeand-Excite modules to stabilize the training in deep networks. The proposed method was applied to publicly available histopathology and chest X-ray databases and achieved superior performance to existing state-of-the-art networks with and without domain adaptation. Depending on the application, our method can improve multi-class classification accuracy by 5-20% compared to DANN introduced in (Ganin, 2014).