Classification of optical coherence tomography images using a capsule network

Abstract

Background: Classification of optical coherence tomography (OCT) images can be achieved with high accuracy using classical convolution neural networks (CNN), a commonly used deep learning network for computer-aided diagnosis. Classical CNN has often been criticized for suppressing positional relations in a pooling layer. Therefore, because capsule networks can learn positional information from images, we attempted application of a capsule network to OCT images to overcome that shortcoming. This study is our attempt to improve classification accuracy by replacing CNN with a capsule network. Methods: From an OCT dataset, we produced a training dataset of 83,484 images and a test dataset of 1000 images. For training, the dataset comprises 37,205 images with choroidal neovascularization (CNV), 11,348 with diabetic macular edema (DME), 8616 with drusen, and 26,315 normal images. The test dataset has 250 images from each category. The proposed model was constructed based on a capsule network for improving classification accuracy. It was trained using the training dataset. Subsequently, the test dataset was used to evaluate the trained model.Results: Classification of OCT images using our method achieved accuracy of 99.6%, which is 3.2 percentage points higher than that of other methods described in the literature. Conclusion: The proposed method achieved classification accuracy results equivalent to those reported for other methods for CNV, DME, drusen, and normal images.

SISE-PC: Semi-supervised Image Subsampling for Explainable Pathology Classification

“…We observe that the proposed subsampled data set that isolates 2% training images around decision boundaries achieves about 96.45% classification accuracy. This performance is comparable to InceptionV3 based model in [2] that achieved 96.1% accuracy with complete training data set. Thus, the proposed sub-sampling method intuitively lies within the 5% sampling size as learned from the random stratified sampling.…”

“…Automated pathology classification has shown to significantly improve patient prioritization and resourcefulness of treatment procedures and patient care [1]. Although deep learning algorithms such as Resnet and InceptionV3 have been established as state-of-the-art [2] for several pathology classification tasks, training these models from scratch can be expensive from the labelled data acquisition and compute resource perspectives. In this work, we present a semi-supervised image sub-sampling method that identifies a minimal subsampled data set that represents the most sample uncertainty in a latent feature space that is encoded using a self-supervised contrastive model [3].…”

“…With approximately 30 million invasive OCT scans being performed each year worldwide, there is a need to identify patients with OCT images that need prioritized attention. The existing work in [2] shows 92-99% classification accuracy for OCT image classification using large annotated OCT image data sets to train a CapsuleNet model. Contrarily, in this work we propose a novel active learning framework [4] to significantly reduce the training complexity by reducing the size of annotated training data set.…”

A Multitask Deep-Learning System to Classify Diabetic Macular Edema for Different Optical Coherence Tomography Devices: A Multicenter Analysis

“…The fourth novel aspect of our DL algorithm is its applicability to three commonly used OCT devices. Previous studies have focused on one or, at most, two commercial OCT devices (15,16,18,20,27,28,32). In this study, we trained CNNs to detect DME using images obtained from three comme-rcial OCT devices, making the screening more generalizable.…”

Narrative review of artificial intelligence in diabetic macular edema: Diagnosis and predicting treatment response using optical coherence tomography

“…In a very recent study by Takumasa et al ,[ 38 ] a novel capsule network has been proposed to address a four-way classification among DME, choroidal neovascular membrane, Drusen, and normal OCT scans. A very large dataset has been used to conduct the experiments, unlike previous approaches where the datasets were very small.…”