How much data is needed to train a medical image deep learning system to achieve necessary high accuracy?

Abstract

The use of Convolutional Neural Networks (CNN) in natural image classification systems has produced very impressive results. Combined with the inherent nature of medical images that make them ideal for deep-learning, further application of such systems to medical image classification holds much promise. However, the usefulness and potential impact of such a system can be completely negated if it does not reach a target accuracy. In this paper, we present a study on determining the optimum size of the training data set necessary to achieve high classification accuracy with low variance in medical image classification systems. The CNN was applied to classify axial Computed Tomography (CT) images into six anatomical classes. We trained the CNN using six different sizes of training data set (5, 10, 20, 50, 100, and 200) and then tested the resulting system with a total of 6000 CT images. All images were acquired from the Massachusetts General Hospital (MGH) Picture Archiving and Communication System (PACS). Using this data, we employ the learning curve approach to predict classification accuracy at a given training sample size. Our research will present a general methodology for determining the training data set size necessary to achieve a certain target classification accuracy that can be easily applied to other problems within such systems.

Cited by in corpus (23)

• Opening the Black Box of Deep Neural Networks via Information
• A scoping review of transfer learning research on medical image analysis using ImageNet
• Learning to detect chest radiographs containing lung nodules using visual attention networks
• Med-BERT: pre-trained contextualized embeddings on large-scale structured electronic health records for disease prediction
• Improving brain computer interface performance by data augmentation with conditional Deep Convolutional Generative Adversarial Networks
• tax2vec: Constructing Interpretable Features from Taxonomies for Short Text Classification
• A Constructive Prediction of the Generalization Error Across Scales
• Scaling Laws for Deep Learning
• Deep Learning in Bioinformatics
• Image Deconvolution via Noise-Tolerant Self-Supervised Inversion
• Hardware-Accelerated SAR Simulation with NVIDIA-RTX Technology
• Understanding the Mechanisms of Deep Transfer Learning for Medical Images
• Data Leverage: A Framework for Empowering the Public in its Relationship with Technology Companies
• Recommending Training Set Sizes for Classification
• Impact of Training Dataset Size on Neural Answer Selection Models
• Lesion Conditional Image Generation for Improved Segmentation of Intracranial Hemorrhage from CT Images
• A Multisite, Report-Based, Centralized Infrastructure for Feedback and Monitoring of Radiology AI/ML Development and Clinical Deployment
• Edge Learning with Unmanned Ground Vehicle: Joint Path, Energy and Sample Size Planning
• How many images do I need? Understanding how sample size per class affects deep learning model performance metrics for balanced designs in autonomous wildlife monitoring
• Slice Tuner: A Selective Data Acquisition Framework for Accurate and Fair Machine Learning Models
• Voxel-level Siamese Representation Learning for Abdominal Multi-Organ Segmentation
• A deep neural network for multi-species fish detection using multiple acoustic cameras
• SUSAN: Segment Unannotated image Structure using Adversarial Network