Fine-Tuning and training of densenet for histopathology image representation using TCGA diagnostic slides

Feature vectors provided by pre-trained deep artificial neural networks have become a dominant source for image representation in recent literature. Their contribution to the performance of image analysis can be improved through fine-tuning. As an ultimate solution, one might even train a deep network from scratch with the domain-relevant images, a highly desirable option which is generally impeded in pathology by lack of labeled images and the computational expense. In this study, we propose a new network, namely KimiaNet, that employs the topology of the DenseNet with four dense blocks, finetuned and trained with histopathology images in different configurations. We used more than 240,000 image patches with 1000 x 1000 pixels acquired at 20x magnification through our proposed high-cellularity mosaic approach to enable the usage of weak labels of 7126 whole slide images of formalinfixed paraffin-embedded human pathology samples publicly available through The Cancer Genome Atlas (TCGA) repository. We tested KimiaNet using three public datasets, namely TCGA, endometrial cancer images, and colorectal cancer images by evaluating the performance of search and classification when corresponding features of different networks are used for image representation. As well, we designed and trained multiple convolutional batch-normalized ReLU (CBR) networks. The results show that KimiaNet provides superior results compared to the original DenseNet and smaller CBR networks when used as feature extractor to represent histopathology images. (c) 2021 The Author(s). Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license ( http://creativecommons.org/licenses/by-nc-nd/4.0/ )

Automated summary

Pre-trained deep artificial neural networks are a dominant source for image representation and their performance in image analysis can be improved through fine-tuning. This study introduces a new network, KimiaNet, which outperforms the original DenseNet and other networks in representing histopathology images. By utilizing a large dataset and training with different configurations, KimiaNet shows superior results in image analysis.