Jian Choong R.ZAustin Harding STang B.-YSHIH-WEI LIAO2021-09-022021-09-0220201557170Xhttps://www.scopus.com/inward/record.uri?eid=2-s2.0-85091031954&doi=10.1109%2fEMBC44109.2020.9175392&partnerID=40&md5=7a2f6e221da39384d983e5fcd85ba70ehttps://scholars.lib.ntu.edu.tw/handle/123456789/581446The difficulty of applying deep learning algorithms to biomedical imaging systems arises from a lack of training images. An existing workaround to the lack of medical training images involves pre-training deep learning models on ImageNet, a non-medical dataset with millions of training images. However, the modality of ImageNet's dataset samples consisting of natural images in RGB frequently differs from the modality of medical images, consisting largely of images in grayscale such as X-ray and MRI scan imaging. While this method may be effectively applied to non-medical tasks such as human face detection, it proves ineffective in many areas of medical imaging. Recently proposed generative models such as Generative Adversarial Networks (GANs) are able to synthesize new medical images. By utilizing generated images, we may overcome the modality gap arising from current transfer learning methods. In this paper, we propose a training pipeline which outperforms both conventional GAN-synthetic methods and transfer learning methods. ? 2020 IEEE.Bioinformatics; Deep learning; Face recognition; Image classification; Learning algorithms; Learning systems; Magnetic resonance imaging; Pipelines; Transfer learning; Adversarial networks; Biomedical imaging; Current transfer; Generative model; Human face detection; Medical training; Synthetic methods; Transfer learning methods; Medical imagingconference paper10.1109/EMBC44109.2020.91753922-s2.0-85091031954