Resolution-based distillation for efficient histology image classification

作者：

Highlights：

• Developed a deep learning model for low-resolution histology image classification

• Performed knowledge distillation to maintain accuracy at low-resolution

• Improved performance further using self-supervision on unlabeled data

• Achieved strong classification performance across multiple datasets and metrics

摘要

•Developed a deep learning model for low-resolution histology image classification•Performed knowledge distillation to maintain accuracy at low-resolution•Improved performance further using self-supervision on unlabeled data•Achieved strong classification performance across multiple datasets and metrics

论文关键词：Deep neural networks,Digital pathology,Knowledge distillation,Self-supervised learning

论文评审过程：Received 11 January 2021, Revised 7 July 2021, Accepted 2 August 2021, Available online 6 August 2021, Version of Record 20 August 2021.

论文官网地址：https://doi.org/10.1016/j.artmed.2021.102136