[Paper Review] Development of a Machine-Learning System to Classify Lung CT Scan Images into Normal/COVID-19 Class

Seifedine Kadry, V. Rajinikanth+4 more|arXiv (Cornell University)|Apr 24, 2020

COVID-19 diagnosis using AIMedicine53 references33 citations

TL;DR

The paper proposes a machine-learning pipeline usingChaotic Bat Algorithm and Kapur’s Entropy thresholding to enhance lung CT slices, followed by feature extraction/selection and a two-class classifier, achieving up to 89.80% accuracy with an SVM using a fused feature vector.

ABSTRACT

Recently, the lung infection due to Coronavirus Disease (COVID-19) affected a large human group worldwide and the assessment of the infection rate in the lung is essential for treatment planning. This research aims to propose a Machine-Learning-System (MLS) to detect the COVID-19 infection using the CT scan Slices (CTS). This MLS implements a sequence of methods, such as multi-thresholding, image separation using threshold filter, feature-extraction, feature-selection, feature-fusion and classification. The initial part implements the Chaotic-Bat-Algorithm and Kapur's Entropy (CBA+KE) thresholding to enhance the CTS. The threshold filter separates the image into two segments based on a chosen threshold 'Th'. The texture features of these images are extracted, refined and selected using the chosen procedures. Finally, a two-class classifier system is implemented to categorize the chosen CTS (n=500 with a pixel dimension of 512x512x1) into normal/COVID-19 group. In this work, the classifiers, such as Naive Bayes (NB), k-Nearest Neighbors (KNN), Decision Tree (DT), Random Forest (RF) and Support Vector Machine with linear kernel (SVM) are implemented and the classification task is performed using various feature vectors. The experimental outcome of the SVM with Fused-Feature-Vector (FFV) helped to attain a detection accuracy of 89.80%.

Motivation & Objective

• Motivate the need for rapid COVID-19 detection from CT scans for treatment planning and triage.
• Propose a machine-learning workflow that enhances CT slices and extracts discriminative texture features.
• Evaluate multiple classifiers to identify effective models for Normal vs COVID-19 classification.

Proposed method

• Apply Chaotic Bat Algorithm and Kapur’s Entropy thresholding to enhance CT slices.
• Perform threshold-based image separation to create two segments at a chosen threshold Th.
• Extract, refine, and select texture features from the segmented images.
• Fuse feature vectors via a fusion scheme and train a two-class classifier.
• Evaluate Naive Bayes, KNN, Decision Tree, Random Forest, and linear SVM with various feature vectors.

Experimental results

Research questions

• RQ1 Can a thresholding-based enhancement combined with texture features effectively distinguish Normal vs COVID-19 CT slices?
• RQ2 Which classifier and feature vector configuration yields the highest accuracy for normal/COVID-19 classification on the given dataset?

Key findings

• The SVM with a fused feature vector achieved 89.80% detection accuracy.
• The study tested multiple classifiers, identifying SVM as the best performer among NB, KNN, DT, RF, and SVM with FFV.
• CT slice dataset used: n=500 with size 512x512x1.