Call us Today ! +918886268863 | [email protected]

JCS: An Explainable COVID-19 Diagnosis System by Joint Classification and Segmentation

JCS: An Explainable COVID-19 Diagnosis System by Joint Classification and Segmentation

Introduction:

CORONAVIRUS disease 2019, or COVID-19, is an epidemic disease caused by the Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2). It outbreaks around the world in a short period and has caused 1,914,916 confirmed cases and 123,010 confirmed deaths as of April 15th, 2020. COVID-19 pushes the health systems of over 200 countries to the brink of collapse due to the lack of medical supplies and staff and thus has been declared as a pandemic by the World Health Organization. The current main diagnostic tool for COVID-19 is via the Reverse Transcription Polymerase Chain Reaction (RT-PCR) test. However, the RT-PCR test is not accurate enough to well prevent the pandemic. So the false-negative cases of RT-PCR tests are a potential threat to public wellness, and missing any COVID-19 cases will probably cause secondary infections of large areas.

Abstracts:

Recently, the coronavirus disease 2019 (COVID-19) has caused a pandemic disease in over 200 countries, influencing billions of humans. To control the infection, identifying and separating the infected people is the most crucial step. The main diagnostic tool is the Reverse Transcription Polymerase Chain Reaction (RT-PCR) test. This paper develops a novel Joint Classification and Segmentation (JCS) system to perform real-time and explainable COVID-19 chest CT diagnosis. To train our JCS system, we construct a large scale COVID-19 Classification and Segmentation (COVID-CS) dataset, with 144,167 chest CT images of 400 COVID-19 patients and 350 uninfected cases. 3,855 chest CT images of 200 patients are annotated with fine-grained pixel-level labels of opacifications, which are increased attenuation of the lung parenchyma. We also have annotated lesion counts, opacification areas, and locations and thus benefit various diagnosis aspects. Extensive experiments demonstrate that the proposed JCS diagnosis system is very efficient for COVID-19 classification and segmentation. It obtains an average sensitivity of 95.0% and a specificity of 93.0% on the classification test set, and 78.5% Dice score on the segmentation test set of our COVID-CS dataset.

Existing work:

As of April 15th, 1,914,916 people are infected by COVID-19. But their CT scans are usually private and could not be publicly accessed. To facilitate the development of diagnostic systems, several COVID-19 related datasets are publicly released by researchers around the world. A summary of these datasets is provided in Table I. One X-ray dataset from Cohen et al. [9] contains overall 122 frontal view X-rays, including 100 images of COVID-19 cases, 11 SARS images, and 11 other pneumonia images. The COVID-CT dataset from [10] has 746 CT scan images, 349 images from COVID-19 patients and 397 from non-COVID-19 cases. All the images in these datasets are collected from public websites and/or COVID-19 related papers on medRxiv, bioRxiv, and journals, etc

Disadvantage:

Still, the sensitivity of the RT-PCR test is not high enough to effectively prevent the pandemic. The chest CT scan test provides a valuable complementary tool to the RTPCR test, and it can identify the patients in the early-stage with high sensitivity. However, the chest CT scan test is usually time consuming, requiring about 21.5 minutes per case.

Proposed work:

This paper develops a novel Joint Classification and Segmentation (JCS) system to perform real-time and explainable COVID-19 chest CT diagnosis. To train our JCS system, we construct a large scale COVID-19 Classification and Segmentation (COVID-CS) dataset, with 144,167 chest CT images of 400 COVID-19 patients and 350 uninfected cases. 3,855 chest CT images of 200 patients are annotated with fine-grained pixel-level labels of opacifications, which are increased attenuation of the lung parenchyma. We also have annotated lesion counts, opacification areas, and locations and thus benefit various diagnosis aspects. Extensive experiments demonstrate that the proposed JCS diagnosis system is very efficient for COVID-19 classification and segmentation.

Advantage:

  • Improve performance
  • Effectiveness

Algorithms: CNN

System requirements:

  Software requirements:

  • Operating system           :   Windows.
  • Coding Language :   Python.

Hardware components:

System                 :   Pentium IV 2.4 GHz or intel

Hard Disk                      :   40 GB.

Floppy Drive       :   1.44 Mb.

Mouse                  :   Optical Mouse.

Ram                      :   512 Mb.

Conclusion:

To facilitate the training of strong CNN models for COVID-19 diagnosis, in this paper, we systematically constructed a large scale COVID-19 Classification and Segmentation (COVID-CS) dataset. We also developed a Joint Classification and Segmentation (JCS) system for COVID-19 diagnosis. In our system, the classification model identified whether the suspected patient is COVID-19 positive or not, along with convincing visual explanations. It obtained a 95.0% sensitivity and 93.0% specificity on the classification test set of our COVID-CS dataset. To provide complementary pixel-level prediction, we implemented a segmentation model to discover fine-grained lesion areas in the CT images of COVID-19 patients. Comparing to the competing methods, e.g., PoolNet [46], our segmentation model achieved an improvement of 8.8% on the Dice metric.

March 19, 2022

0 responses on "JCS: An Explainable COVID-19 Diagnosis System by Joint Classification and Segmentation"

Leave a Message

Template Design © VibeThemes. All rights reserved.