COVID19 Diagnosis using AutoML from 3D CT scans

Anwar, Talha

doi:10.1109/iccvw54120.2021.00061

Cited by 15 publications

(8 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Among these techniques, segmentation of the lung region using image pre-processing methods is applied before performing classification. Owing to the varying dimensionality of 3D CT scans, interpolation or truncation of the slices are applied to convert them to fixed dimensionality which might lead to information loss [41]. In 2D CT scan-based classification, a 2D CNN is trained on individual slices and slice-level probability scores are generated.…”

Section: B Ct Scan Classificationmentioning

confidence: 99%

“…In 2D CT scan-based classification, a 2D CNN is trained on individual slices and slice-level probability scores are generated. Further, thresholdbased [41], majority voting [42], [43], weighted average methods [44] and sequence models (such as recurrent neural network (RNN) [45] and bidirectional long short term memory (BiLSTM) [46]) are used to obtain patient-level COVID-19 classification. Threshold-based and majority voting methods create higher false negatives at regions, where the traces of infections are not visible as in the majority of CT scan slices.…”

Section: B Ct Scan Classificationmentioning

confidence: 99%

See 1 more Smart Citation

Identification and Severity Assessment of COVID-19 Using Lung CT Scans

Thyagachandran¹,

Balachandran²,

Murthy³

2023

Preprint

View full text Add to dashboard Cite

<p>The COVID-19 pandemic, caused by the SARS-CoV-2 virus continues to have a significant impact on the global population. To effectively triage patients and understand the progression of the disease, a metric-based analysis of diagnostic techniques is necessary. The objective of the present study is to identify COVID-19 from chest CT scans and determine the extent of severity, defined by a severity score that indicates the volume of infection. An unsupervised preprocessing pipeline is proposed to extract relevant clinical features and utilize this information to employ a pre-trained ImageNet model to extract discriminative features. Subsequently, a shallow feed-forward neural network is trained to classify the available CT scans into three classes namely COVID-19, Community-Acquired Pneumonia, and Normal. Through various ablation studies, we find that a domain-specific pre-processing pipeline improves classification accuracy significantly. In terms of classification accuracy, our approach when evaluated on publicly available datasets is seen to have an absolute improvement of 6\% F1 score over the baseline model. Further, the estimated infection severity score is observed to be well correlated with radiologists' assessments. The results support the necessity of data-driven pre-processing before implementing learning algorithms.</p>

show abstract

Section: B Ct Scan Classificationmentioning

confidence: 99%

Section: B Ct Scan Classificationmentioning

confidence: 99%

Identification and Severity Assessment of COVID-19 Using Lung CT Scans

Thyagachandran¹,

Balachandran²,

Murthy³

2023

Preprint

View full text Add to dashboard Cite

show abstract

“…Even though a variety of machine-guided automated labeling [12], [53], [119], [120], [121] techniques have been proposed, they are generally limited to simple labeling tasks targeting very narrow application settings. It is currently challenging to extend them to more complex or general tasks.…”

Section: Automated Data Labelingmentioning

confidence: 99%

Data augmentation: A comprehensive survey of modern approaches

Mumuni

2022

Array

250

View full text Add to dashboard Cite

“…The calibration block is a 3D 1×1×1 point-wise convolution to solve the problem of feature dimension mismatch; thus, all subsequent blocks have a stride of 1. The number of searchable blocks and the stride of calibration block in six layers are [4,4,4,4,4,1] and [2,2,2,1,2,1], respectively. The output channels of the stem block and six layers are 32 and [24,40,80,96,192,320], respectively.…”

Section: Search Spacementioning

confidence: 99%

Evolutionary Multi-objective Architecture Search Framework: Application to COVID-19 3D CT Classification

Ying

Zhang

et al. 2022

Lecture Notes in Computer Science

View full text Add to dashboard Cite

The COVID-19 pandemic has threatened global health. Many studies have applied deep convolutional neural networks (CNN) to recognize COVID-19 based on chest 3D computed tomography (CT). Recent works show that no model generalizes well across CT datasets from different countries, and manually designing models for specific datasets requires expertise; thus, neural architecture search (NAS) that aims to search models automatically has become an attractive solution. To reduce the search cost on large 3D CT datasets, most NAS-based works use the weight-sharing (WS) strategy to make all models share weights within a supernet; however, WS inevitably incurs search instability, leading to inaccurate model estimation. In this work, we propose an efficient Evolutionary Multi-objective ARchitecture Search (EMARS) framework. We propose a new objective, namely potential, which can help exploit promising models to indirectly reduce the number of models involved in weights training, thus alleviating search instability. We demonstrate that under objectives of accuracy and potential, EMARS can balance exploitation and exploration, i.e., reducing search time and finding better models. Our searched models are small and perform better than prior works on three public COVID-19 3D CT datasets.

show abstract

COVID19 Diagnosis using AutoML from 3D CT scans

Cited by 15 publications

References 18 publications

Identification and Severity Assessment of COVID-19 Using Lung CT Scans

Identification and Severity Assessment of COVID-19 Using Lung CT Scans

Data augmentation: A comprehensive survey of modern approaches

Evolutionary Multi-objective Architecture Search Framework: Application to COVID-19 3D CT Classification

Contact Info

Product

Resources

About