Deep Convolutional Neural Networks for Computer-Aided Detection: CNN Architectures, Dataset Characteristics and Transfer Learning

Abstract

Remarkable progress has been made in image recognition, primarily due to the availability of large-scale annotated datasets (i.e. ImageNet) and the revival of deep convolutional neural networks (CNN). CNNs enable learning data-driven, highly representative, layered hierarchical image features from sufficient training data. However, obtaining datasets as comprehensively annotated as ImageNet in the medical imaging domain remains a challenge. There are currently three major techniques that successfully employ CNNs to medical image classification: training the CNN from scratch, using off-the-shelf pre-trained CNN features, and conducting unsupervised CNN pre-training with supervised fine-tuning. Another effective method is transfer learning, i.e., fine-tuning CNN models (supervised) pre-trained from natural image dataset to medical image tasks (although domain transfer between two medical image datasets is also possible). In this paper, we exploit three important, but previously understudied factors of employing deep convolutional neural networks to computer-aided detection problems. We first explore and evaluate different CNN architectures. The studied models contain 5 thousand to 160 million parameters, and vary in numbers of layers. We then evaluate the influence of dataset scale and spatial image context on performance. Finally, we examine when and why transfer learning from pre-trained ImageNet (via fine-tuning) can be useful. We study two specific computeraided detection (CADe) problems, namely thoraco-abdominal lymph node (LN) detection and interstitial lung disease (ILD) classification. We achieve the state-of-the-art performance on the mediastinal LN detection, with 85% sensitivity at 3 false positive per patient, and report the first five-fold cross-validation classification results on predicting axial CT slices with ILD categories. Our extensive empirical evaluation, CNN model analysis and valuable insights can be extended to the design of high performance CAD systems for other medical imaging tasks.

Learning laparoscopic video shot classification for gynecological surgery

“…Furthermore, there haven't been any efforts made regarding the classification of images extracted from laparoscopic surgery videos. Fine tuning and transfer learning effects of CNNs are covered in recent literature by Shin et al [24] as well as Tajbakhsh et al [26]. These pieces of work are based on the use cases of lymph node detection, interstitial lung disease classification, polyp detection and image quality assessment in colonoscopy, pulmonary embolism detection in computed tomography images, and intima-media boundary segmentation in ultrasonographic images.…”

Respiratory magnetic resonance imaging biomarkers in Duchenne muscular dystrophy

“…The initial network structure was based on a VGGNet model,19 pretrained on ImageNet,20 and then fine‐tuned (as in ref. 21) on a randomly selected frame of each cine breathing MRI dataset with a learning rate of 10 −6 . The trained model was subsequently applied to the entire set of frame sequences (typically 180 frames) in each dataset to define the contours of the lung and the diaphragm.…”

Image Quality Classification for DR Screening Using Convolutional Neural Networks

“…There success is mainly attributed to faster processing (GPUs), rectified linear units, dropout regularisation, and effective data augmentation [13]. Currently there are three main techniques used in training CNN's for medical imaging applications [13]: training the CNN from scratch, use trained off-the-shelf CNN's (i.e. AlexNet [8], VGG-Net [14]) to extract features, unsupervised pre training on large image datasets.…”