[Paper Review] COVID_MTNet: COVID-19 Detection with Multi-Task Deep Learning Approaches

Md Zahangir Alom, M M Shaifur Rahman+3 more|arXiv (Cornell University)|Apr 7, 2020

COVID-19 diagnosis using AI83 citations

TL;DR

This paper presents a multi-task deep learning framework for COVID-19 detection from X-ray and CT images, using an Inception Residual Recurrent CNN with Transfer Learning for detection and NABLA-N for infected-region segmentation, plus a new quantitative analysis for infection percentages.

ABSTRACT

COVID-19 is currently one the most life-threatening problems around the world. The fast and accurate detection of the COVID-19 infection is essential to identify, take better decisions and ensure treatment for the patients which will help save their lives. In this paper, we propose a fast and efficient way to identify COVID-19 patients with multi-task deep learning (DL) methods. Both X-ray and CT scan images are considered to evaluate the proposed technique. We employ our Inception Residual Recurrent Convolutional Neural Network with Transfer Learning (TL) approach for COVID-19 detection and our NABLA-N network model for segmenting the regions infected by COVID-19. The detection model shows around 84.67% testing accuracy from X-ray images and 98.78% accuracy in CT-images. A novel quantitative analysis strategy is also proposed in this paper to determine the percentage of infected regions in X-ray and CT images. The qualitative and quantitative results demonstrate promising results for COVID-19 detection and infected region localization.

Motivation & Objective

• Motivate fast and accurate COVID-19 detection to aid clinical decisions and treatment planning.
• Develop a multi-task deep learning pipeline that handles both detection and infection-region segmentation.
• Leverage transfer learning to improve detection performance on medical imaging datasets.
• Provide a quantitative method to estimate the percentage of infected regions in images.

Proposed method

• Employ an Inception Residual Recurrent Convolutional Neural Network with Transfer Learning for COVID-19 detection.
• Apply the NABLA-N network for segmenting regions infected by COVID-19.
• Propose a novel quantitative analysis strategy to determine the percentage of infected regions in X-ray and CT images.
• Evaluate the approach on both X-ray and CT image datasets to assess robustness across modalities.

Experimental results

Research questions

• RQ1 How accurately can multi-task deep learning detect COVID-19 from X-ray and CT images?
• RQ2 Can segmentation of infected regions complement detection to improve overall diagnostic utility?
• RQ3 Does transfer learning improve detection performance on limited COVID-19 imaging data?
• RQ4 What is the proposed method for quantifying infected-region percentage in images?

Key findings

• Detection accuracy of approximately 84.67% on X-ray images.
• Detection accuracy of approximately 98.78% on CT images.
• A multi-task framework combining detection and infection-region segmentation was effective.
• A novel quantitative analysis strategy estimates the percentage of infected regions in both X-ray and CT images.