Deep learning in generating radiology reports: A survey
作者：
Highlights：
• Deep Learning algorithms showed promising results in generating radiology reports.
• We categorize state of the art models into three levels: word, sentence and paragraph.
• We review the publicly available datasets of radiology images and linked reports.
• We compare results of the generated reports through quantitative evaluation matrices.
• Researchers integrate convolutional neural network and recurrent neural network.
摘要
•Deep Learning algorithms showed promising results in generating radiology reports.•We categorize state of the art models into three levels: word, sentence and paragraph.•We review the publicly available datasets of radiology images and linked reports.•We compare results of the generated reports through quantitative evaluation matrices.•Researchers integrate convolutional neural network and recurrent neural network.
论文关键词：Convolutional neural network,Deep learning,Natural language processing,Radiology,Recurrent neural network
论文评审过程：Received 16 April 2019, Revised 30 April 2020, Accepted 10 May 2020, Available online 15 May 2020, Version of Record 23 May 2020.
论文官网地址：https://doi.org/10.1016/j.artmed.2020.101878
