A Mask-guided Attention Deep Learning Model for COVID-19 Diagnosis based on an Integrated CT Scan Images Database

Abstract

<p>The global extent of COVID-19 mutations and the consequent depletion of hospital resources highlighted the necessity of effective computer-assisted medical diagnosis. COVID-19 detection mediated by deep learning models can help diagnose this highly contagious disease and lower infectivity and mortality rates. Computed tomography (CT) is the preferred imaging modality for building automatic COVID-19 screening and diagnosis models. It is well-known that the training set size significantly impacts the performance and generalization of deep learning models. However, accessing a large dataset of CT scan images from an emerging disease like COVID-19 is challenging. Therefore, data efficiency becomes a significant factor in choosing a learning model. To this end, we present a multi-task learning approach, namely, a mask-guided attention (MGA) classifier, to improve the generalization and data efficiency of COVID-19 classification on lung CT scan images.</p><p>The novelty of this method is compensating for the scarcity of data by employing more supervision with lesion masks, increasing the sensitivity of the model to COVID-19 manifestations, and helping both generalization and classification performance. Our proposed model achieves better overall performance than the single-task baseline and state-of-the-art models, as measured by various popular metrics. In our experiment with different percentages of data from our curated dataset, the classification performance gain from this multi-task learning approach is more significant for the smaller training sizes. Furthermore, experimental results demonstrate that our method enhances the focus on the lesions, as witnessed by both</p><p>attention and attribution maps, resulting in a more interpretable model.</p>

TSRNet: Diagnosis of COVID-19 based on self-supervised learning and hybrid ensemble model

“…(b) We pre-trained RANet on the ImageNet dataset and then fine-tuned it accordingly on the target dataset. (c) We pre-trained RANet on the [ 39 ] dataset and then fine-tuned it accordingly in the downstream task. (d) We first pre-trained the RANet network on the ImageNet dataset, then transferred it to the [ 39 ] dataset for training, and finally fine-tuned the network on the target dataset.…”

“…(c) We pre-trained RANet on the [ 39 ] dataset and then fine-tuned it accordingly in the downstream task. (d) We first pre-trained the RANet network on the ImageNet dataset, then transferred it to the [ 39 ] dataset for training, and finally fine-tuned the network on the target dataset. The experimental results are shown in Table 4 .…”

“…11 . In addition, since the ImageNet dataset is more different from the target dataset in terms of number and type, the [ 39 ] dataset is closer to the target dataset.…”

“…Therefore, transfer learning performance is better when the source domain data is closer to the target domain data. Finally, the [ 39 ] dataset has a single data type and small data volume. Therefore, after pre-training the network model on the large dataset, transferring it to a smaller dataset closer to the target domain data, and pre-training it again, the network will learn more unbiased information about the image representation.…”

“…Finally, to verify the impact of the difference between the source domain and the target domain on transfer learning, the data in Ref. [ 39 ] was used as a secondary dataset for validation.

Fig.

…”