Active, Continual Fine Tuning of Convolutional Neural Networks for Reducing Annotation Efforts

Abstract

The splendid success of convolutional neural networks (CNNs) in computer vision is largely attributable to the availability of massive annotated datasets, such as ImageNet and Places. However, in medical imaging, it is challenging to create such large annotated datasets, as annotating medical images is not only tedious, laborious, and time consuming, but it also demands costly, specialty-oriented skills, which are not easily accessible. To dramatically reduce annotation cost, this paper presents a novel method to naturally integrate active learning and transfer learning (fine-tuning) into a single framework, which starts directly with a pre-trained CNN to seek "worthy" samples for annotation and gradually enhances the (fine-tuned) CNN via continual fine-tuning. We have evaluated our method using three distinct medical imaging applications, demonstrating that it can reduce annotation efforts by at least half compared with random selection.

References in corpus (26)

• Very Deep Convolutional Networks for Large-Scale Image Recognition
• A Survey on Deep Learning in Medical Image Analysis
• Automated Design of Deep Learning Methods for Biomedical Image Segmentation
• Overcoming catastrophic forgetting in neural networks
• A Simple Framework for Contrastive Learning of Visual Representations
• Convolutional Neural Networks for Medical Image Analysis: Full Training or Fine Tuning?
• A systematic study of the class imbalance problem in convolutional neural networks
• Domain-adversarial neural networks to address the appearance variability of histopathology images
• A Survey on Active Learning and Human-in-the-Loop Deep Learning for Medical Image Analysis
• Deep Bayesian Active Learning with Image Data
• Med3D: Transfer Learning for 3D Medical Image Analysis
• Embracing Imperfect Datasets: A Review of Deep Learning Solutions for Medical Image Segmentation
• Models Genesis
• Truncated Power Method for Sparse Eigenvalue Problems
• Semi-Supervised Medical Image Segmentation via Learning Consistency under Transformations
• Learning Active Learning from Data
• Demystifying Contrastive Self-Supervised Learning: Invariances, Augmentations and Dataset Biases
• Uncertainty-aware Self-ensembling Model for Semi-supervised 3D Left Atrium Segmentation
• Models Genesis: Generic Autodidactic Models for 3D Medical Image Analysis
• Dropout-based Active Learning for Regression
• Transformation Consistent Self-ensembling Model for Semi-supervised Medical Image Segmentation
• Active Learning for Segmentation by Optimizing Content Information for Maximal Entropy
• Semi-Supervised Brain Lesion Segmentation with an Adapted Mean Teacher Model
• Cost-Sensitive Active Learning for Intracranial Hemorrhage Detection
• Learning Semantics-enriched Representation via Self-discovery, Self-classification, and Self-restoration
• Extreme Consistency: Overcoming Annotation Scarcity and Domain Shifts

Cited by in corpus (3)

• A Survey on Active Learning and Human-in-the-Loop Deep Learning for Medical Image Analysis
• Multistep Automated Data Labelling Procedure (MADLaP) for Thyroid Nodules on Ultrasound: An Artificial Intelligence Approach for Automating Image Annotation
• High-Dimensional Data Set Simplification by Laplace-Beltrami Operator