Human-level COVID-19 Diagnosis from Low-dose CT Scans Using a Two-stage Time-distributed Capsule Network

Abstract

Reverse transcription-polymerase chain reaction (RT-PCR) is currently the gold standard in COVID-19 diagnosis. It can, however, take days to provide the diagnosis, and false negative rate is relatively high. Imaging, in particular chest computed tomography (CT), can assist with diagnosis and assessment of this disease. Nevertheless, it is shown that standard dose CT scan gives significant radiation burden to patients, especially those in need of multiple scans. In this study, we consider low-dose and ultra-low-dose (LDCT and ULDCT) scan protocols that reduce the radiation exposure close to that of a single X-Ray, while maintaining an acceptable resolution for diagnosis purposes. Since thoracic radiology expertise may not be widely available during the pandemic, we develop an Artificial Intelligence (AI)-based framework using a collected dataset of LDCT/ULDCT scans, to study the hypothesis that the AI model can provide human-level performance. The AI model uses a two stage capsule network architecture and can rapidly classify COVID-19, community acquired pneumonia (CAP), and normal cases, using LDCT/ULDCT scans. The AI model achieves COVID-19 sensitivity of 89.5% ± 0.11, CAP sensitivity of 95% ± 0.11, normal cases sensitivity (specificity) of 85.7% ± 0.16, and accuracy of 90% ± 0.06. By incorporating clinical data (demographic and symptoms), the performance further improves to COVID-19 sensitivity of 94.3% ± 0.05, CAP sensitivity of 96.7% ± 0.07, normal cases sensitivity (specificity) of 91% ± 0.09 , and accuracy of 94.1% ± 0.03. The proposed AI model achieves human-level diagnosis based on the LDCT/ULDCT scans with reduced radiation exposure. We believe that the proposed AI model has the potential to assist the radiologists to accurately and promptly diagnose COVID-19 infection and help control the transmission chain during the pandemic.

Diabetes Noninvasive Recognition via Improved Capsule Network

“…In [22], the variation of CapsNet named spectral capsule network is proposed, which improves the stability of CapsNet with EM routing. CapsNet has also been used in various tasks and fields, such as iris recognition [23], classification of breast cancer [24]- [26], hyperspectral image spectral classification [27]- [29], gait recognition [30], [31] and pathological lungs segmentation from low dose CT scans [32], [33].…”