Multi-Stage Classification of Retinal OCT Using Multi-Scale Ensemble Deep Architecture

Abstract

Accurate noninvasive diagnosis of retinal disorders is required for appropriate treatment or precision medicine. This work proposes a multi-stage classification network built on a multi-scale (pyramidal) feature ensemble architecture for retinal image classification using optical coherence tomography (OCT) images. First, a scale-adaptive neural network is developed to produce multi-scale inputs for feature extraction and ensemble learning. The larger input sizes yield more global information, while the smaller input sizes focus on local details. Then, a feature-rich pyramidal architecture is designed to extract multi-scale features as inputs using DenseNet as the backbone. The advantage of the hierarchical structure is that it allows the system to extract multi-scale, information-rich features for the accurate classification of retinal disorders. Evaluation on two public OCT datasets containing normal and abnormal retinas (e.g., diabetic macular edema (DME), choroidal neovascularization (CNV), age-related macular degeneration (AMD), and Drusen) and comparison against recent networks demonstrates the advantages of the proposed architecture’s ability to produce feature-rich classification with average accuracy of 97.78%, 96.83%, and 94.26% for the first (binary) stage, second (three-class) stage, and all-at-once (four-class) classification, respectively, using cross-validation experiments using the first dataset. In the second dataset, our system showed an overall accuracy, sensitivity, and specificity of 99.69%, 99.71%, and 99.87%, respectively. Overall, the tangible advantages of the proposed network for enhanced feature learning might be used in various medical image classification tasks where scale-invariant features are crucial for precise diagnosis.

Stitched vision transformer for age-related macular degeneration detection using retinal optical coherence tomography images

“…Even though fusing multi-scale features allows the model to use both high and low-level features, the time and cost of model learning and inference increases compared to classical CNN models due to several different CNN structures, from shallow to deep. In [ 42 ], using a scale-adaptive network, multi-scale features were extracted from the input images, and further, with the help of fusing the extracted features in a hierarchical structure, the classification accuracy on UCSD data was improved to 99.69% in the five-fold cross-validation method. Their multi-task model exhibited competitive results relative to previous state-of-the-art models.…”

Toward Smart, Automated Junctional Tourniquets—AI Models to Interpret Vessel Occlusion at Physiological Pressure Points

“…We hypothesized that the skill threshold could be overcome through the use of artificial intelligence (AI) models that can detect the appropriate vessel requiring occlusion from real-time ultrasound video feeds. AI has begun to revolutionize medicine through smart, precision-medicine applications such as categorizing a wide range of abnormal states from optical coherence tomography scans [22], compiling large diverse data sets for making medical decisions [23], and using predictive text AI models to provide medical recommendations during telemedicine [24]. Smart medicine applications have been extensively reviewed elsewhere [25][26][27][28][29][30].…”

Explainable ensemble learning method for OCT detection with transfer learning

“…On the dataset from UCSD, FN-OCT achieved 98.7% accuracy and 99.1% AUC, surpassing InceptionV3 by 5.3%. Akinniyi et al [35] proposed a multi-stage classification network using OCT images for retinal image classification. Their architecture utilizes a pyramidal feature ensemble built on Dense-Net for extracting multi-scale features.…”