Data Shapley Value for Handling Noisy Labels: An Application in Screening Covid-19 Pneumonia from Chest Ct Scans

A long-standing challenge of deep learning models involves how to handle noisy labels, especially in applications where human lives are at stake. Adoption of the data Shapley Value (SV), a cooperative game-theoretic approach, is an intelligent valuation solution to tackle the issue of noisy labels. Data SV can be used together with a learning model and an evaluation metric to validate each training point's contribution to the model's performance. The SV of a data point, however, is not unique and depends on the learning model, the evaluation metric, and other data points collaborating in the training game. However, effects of utilizing different evaluation metrics for computation of the SV, detecting the noisy labels, and measuring the data points' importance has not yet been thoroughly investigated. In this context, we performed a series of comparative analyses to assess SV's capabilities to detect noisy input labels when measured by different evaluation metrics. Our experiments on COVID-19-infected of CT images illustrate that although the data SV can effectively identify noisy labels, adoption of different evaluation metric can significantly influence its ability to identify noisy labels from different data classes. Specifically, we demonstrate that the SV greatly depends on the associated evaluation metric.

A Generalization Enhancement Approach for Deep Learning Segmentation Models: Application in COVID-19 Lesion Segmentation from Chest CT Slices

Deep learning has shown remarkable promise in medical imaging tasks, reaching an expert level of performance for some diseases. However, these models often fail to generalize properly to data not used during training, which is a major roadblock to successful clinical deployment. This paper proposes a generalization enhancement approach that can mitigate the gap between source and unseen data in deep learning-based segmentation models without using ground-truth masks of the target domain. Leveraging a subset of unseen domain's CT slices for which the model trained on the source data yields the most confident predictions and their predicted masks, the model learns helpful features of the unseen data over a retraining process. We investigated the effectiveness of the introduced method over three rounds of experiments on three open-access COVID-19 lesion segmentation datasets, and the results illustrate constant improvements of the segmentation model performance on datasets not seen during training. © 2022 European Signal Processing Conference, EUSIPCO. All rights reserved.

Wso-caps: Diagnosis of lung infection from low and ultra-lowdose CT scans using capsule networks and windowsetting optimization

The automatic diagnosis of lung infections using chest computed tomography (CT) scans has been recently obtained remarkable significance, particularly during the COVID-19 pandemic that the early diagnosis of the disease is of utmost importance. In addition, infection diagnosis is the main building block of most automated diagnostic/prognostic frameworks. Recently, due to the devastating effects of the radiation on the body caused by the CT scan, there has been a surge in acquiring low and ultra-low-dose CT scans instead of the standard scans. Such CT scans, however, suffer from a high noise level which makes them difficult and time-consuming to interpret even by expert radiologists. In addition, some abnormalities are only visible using specific window settings on the radiologists' monitor. Currently, manual adjustment of the windowing settings is the common approach to analyze such low-quality images. In this paper, we propose an automated framework based on the Capsule Networks, referred to as the 'WSO-CAPS', to detect slices demonstrating infection using low and ultra-low-dose chest CT scans. The WSOCAPS framework is equipped with a Window Setting Optimization (WSO) mechanism to automatically identify the best window setting parameters to resemble the radiologists' efforts. The experimental results on our in-house dataset show that the WSO-CAPS enhances the capability of the Capsule Network and its counterparts to identify slices demonstrating infection. The WSO-CAPS achieves the accuracy of 92.0%, sensitivity of 90.3%, and specificity of 93.3%. We believe that the proposed WSO-CAPS has a high potential to be further utilized in future frameworks that are working with CT scans, particularly the ones which utilize an infection diagnosis step in their pipeline. © 2021 IEEE.

An ensemble learning framework for multi-class Covid-19 lesion segmentation from chest Ct images

The novel Coronavirus disease (COVID-19) has been the most critical global challenge over the past months. Lung involvement quantification and distinguishing the types of infections from chest CT scans can assist in accurate severity assessment of COVID-19 pneumonia, efficient use of limited medical resources, and saving more lives. Nevertheless, visual assessment of chest CT images and evaluating the disease severity by radiologists are expensive and prone to error. This paper proposes an automated deep learning (DL)-based framework for multi-class segmentation of COVID lesions from chest CT images that takes the CT images as the input and generates a mask indicating the infection regions. The infection regions are segmented under two classes of data, GGOs and consolidations, which are the most common CT patterns of COVID-19 pneumonia. The proposed end-to-end framework contains four encoder-decoder-based segmentation networks that exploit the top-performing pretrained CNNs as the encoder paths and are developed and trained separately. The results then are aggregated using a pixel-level Soft Majority Voting to obtain the final class membership probabilities for each pixel of the image. The proposed framework is evaluated using an open-access CT segmentation dataset. The experimental results indicate that our method successfully performs multi-class segmenting of COVID-19 lung infection regions and outperforms previous works. © 2021 IEEE.

COVID-CT-MD, COVID-19 computed tomography scan dataset applicable in machine learning and deep learning

Novel Coronavirus (COVID-19) has drastically overwhelmed more than 200 countries affecting millions and claiming almost 2 million lives, since its emergence in late 2019. This highly contagious disease can easily spread, and if not controlled in a timely fashion, can rapidly incapacitate healthcare systems. The current standard diagnosis method, the Reverse Transcription Polymerase Chain Reaction (RT- PCR), is time consuming, and subject to low sensitivity. Chest Radiograph (CXR), the first imaging modality to be used, is readily available and gives immediate results. However, it has notoriously lower sensitivity than Computed Tomography (CT), which can be used efficiently to complement other diagnostic methods. This paper introduces a new COVID-19 CT scan dataset, referred to as COVID-CT-MD, consisting of not only COVID-19 cases, but also healthy and participants infected by Community Acquired Pneumonia (CAP). COVID-CT-MD dataset, which is accompanied with lobe-level, slice-level and patient-level labels, has the potential to facilitate the COVID-19 research, in particular COVID-CT-MD can assist in development of advanced Machine Learning (ML) and Deep Neural Network (DNN) based solutions.