Evaluating histopathology transfer learning with ChampKit

  title={Evaluating histopathology transfer learning with ChampKit},
  author={Jakub R. Kaczmarzyk and Tahsin M. Kurç and Shahira Abousamra and Rajarsi R. Gupta and Joel H. Saltz and Peter K. Koo},
Histopathology remains the gold standard for diagnosis of various cancers. Recent 1 advances in computer vision, specifically deep learning, have facilitated the anal- 2 ysis of histopathology images for various tasks, including immune cell detection 3 and microsatellite instability classification. The state-of-the-art for each task often 4 employs base architectures that have been pretrained for image classification on 5 ImageNet. The standard approach to develop classifiers in histopathology… 
1 Citations

Figures and Tables from this paper

Global Research Trends of Artificial Intelligence on Histopathological Images: A 20-Year Bibliometric Analysis

The first bibliometric analysis of AI in HI from multiple indicators is performed, providing insights for researchers to identify key cancer types and understand the research trends of AI application in HI.



HistoTransfer: Understanding Transfer Learning for Histopathology

This work compares the performance of features extracted from networks trained on ImageNet and histopathology data, and uses an attention pooling network over these extracted features for slide-level aggregation to investigate if features learned using more complex networks lead to gain in performance.

Classification of Histopathological Biopsy Images Using Ensemble of Deep Learning Networks

An ensemble deep learning-based approach for automatic binary classification of breast histology images by adapting three pre-trained CNNs, namely VGG19, MobileNet, and DenseNet, which obtains better predictions than single classifiers and machine learning algorithms.

Breast Cancer Histopathology Image Classification Using an Ensemble of Deep Learning Models

An ensemble deep learning approach is presented for the definite classification of non-carcinoma and carcinoma breast cancer histopathology images using the authors' collected dataset and demonstrated that the proposed deepLearning approach is effective for the automatic classification of complex-natured histopathological images of breast cancer.

Deep Learning for Classification of Colorectal Polyps on Whole-slide Images

An automatic image analysis method that can accurately classify different types of colorectal polyps on whole-slide images to help pathologists with this characterization and diagnosis and can reduce the cognitive burden on pathologists and improve their efficacy in histopathological characterization and follow-up recommendations.

From CNNs to Vision Transformers -- A Comprehensive Evaluation of Deep Learning Models for Histopathology

An extensive evaluation of state-of-the-art models based on essential but complementary quality requirements beyond a mere classification accuracy resulted in model recommendations for practitioners as well as putting forward a general methodology to quantify a model’s quality according to complementary requirements that can be transferred to future model architectures.

Deep learning for magnification independent breast cancer histopathology image classification

Experimental results show that the magnification independent CNN approach improved the performance of magnification specific model, and the results in this limited set of training data are comparable with previous state-of-the-art results obtained by hand-crafted features.

Deep-Learning–Based Characterization of Tumor-Infiltrating Lymphocytes in Breast Cancers From Histopathology Images and Multiomics Data

An automatic pipeline based on a cascade-training U-net to generate high-resolution TIL maps on WSIs can facilitate computational pathology-based discovery in cancer immunology and research on immunotherapy.

Classification and mutation prediction from non–small cell lung cancer histopathology images using deep learning

A deep convolutional neural network model is trained on whole-slide images obtained from The Cancer Genome Atlas to accurately and automatically classify them into LUAD, LUSC or normal lung tissue and predicts the ten most commonly mutated genes in LUAD.