TY - JOUR
T1 - Fine-Tuning and training of densenet for histopathology image representation using TCGA diagnostic slides
AU - Riasatian, Abtin
AU - Babaie, Morteza
AU - Maleki, Danial
AU - Kalra, Shivam
AU - Valipour, Mojtaba
AU - Hemati, Sobhan
AU - Zaveri, Manit
AU - Safarpoor, Amir
AU - Shafiei, Sobhan
AU - Afshari, Mehdi
AU - Rasoolijaberi, Maral
AU - Sikaroudi, Milad
AU - Adnan, Mohd
AU - Shah, Sultaan
AU - Choi, Charles
AU - Damaskinos, Savvas
AU - Campbell, Clinton JV
AU - Diamandis, Phedias
AU - Pantanowitz, Liron
AU - Kashani, Hany
AU - Ghodsi, Ali
AU - Tizhoosh, H. R.
N1 - Publisher Copyright:
© 2021 The Author(s)
PY - 2021/5
Y1 - 2021/5
N2 - 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, fine-tuned and trained with histopathology images in different configurations. We used more than 240,000 image patches with 1000×1000 pixels acquired at 20× magnification through our proposed “high-cellularity mosaic” approach to enable the usage of weak labels of 7126 whole slide images of formalin-fixed 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.
AB - 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, fine-tuned and trained with histopathology images in different configurations. We used more than 240,000 image patches with 1000×1000 pixels acquired at 20× magnification through our proposed “high-cellularity mosaic” approach to enable the usage of weak labels of 7126 whole slide images of formalin-fixed 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.
KW - Deep features
KW - Deep learning
KW - Histopathology
KW - Image classification
KW - Image representation
KW - Image search
KW - TCGA
KW - Transfer learning
UR - http://www.scopus.com/inward/record.url?scp=85103138920&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85103138920&partnerID=8YFLogxK
U2 - 10.1016/j.media.2021.102032
DO - 10.1016/j.media.2021.102032
M3 - Article
C2 - 33773296
AN - SCOPUS:85103138920
SN - 1361-8415
VL - 70
JO - Medical Image Analysis
JF - Medical Image Analysis
M1 - 102032
ER -