COVID-19 Detection Using Feature Extraction and Semi-Supervised Learning from Chest X-ray Images

The lack of labeled data is one of the main prohibiting issues on the development of deep learning models, as they rely on large labeled datasets in order to achieve high accuracy in complex tasks. Our objective is to evaluate the performance gain of having additional unlabeled data in the training of a deep learning model when working with medical imaging data. We present a semi-supervised learning algorithm that utilizes a teacher-student paradigm in order to leverage unlabeled data in the classification of chest X-ray images. Using our algorithm on the ChestX-ray14 dataset, we manage to achieve a substantial increase in performance when using small labeled datasets. With our method, a model achieves an AUROC of 0.822 with only 2% labeled data and 0.865 with 5% labeled data, while a fully supervised method achieves an AUROC of 0.807 with 5% labeled data and only 0.845 with 10%.

Download Full-text

A Novel Machine Learning Approach for Tuberculosis Segmentation and Prediction Using Chest-X-Ray (CXR) Images

Applied Sciences ◽

10.3390/app11199057 ◽

2021 ◽

Vol 11 (19) ◽

pp. 9057

Author(s):

Xavier Alphonse Inbaraj ◽

Charlyn Villavicencio ◽

Julio Jerison Macrohon ◽

Jyh-Horng Jeng ◽

Jer-Guang Hsieh

Keyword(s):

Feature Extraction ◽

Decision Tree ◽

High Morbidity ◽

Decision Tree Classifier ◽

X Ray ◽

Poor Countries ◽

Tree Classifier ◽

Machine Learning Approach ◽

Chest X Ray ◽

Novel Method

Tuberculosis is a potential fatal disease with high morbidity and mortality rates. Tuberculosis death rates are rising, posing a serious health threat in several poor countries around the world. To address this issue, we proposed a novel method for detecting tuberculosis in chest X-ray (CXR) images that uses a three-phased approach to distinguish tuberculosis such as segmentation, feature extraction, and classification. In a CXR, we utilized the Weiner filter to distinguish and reduce the impulse noise. The features were extracted from CXR images and trained using a decision tree classifier known as the stacked loopy decision tree (SLDT) classifier. For the classification process, the ROI-based morphological approach was applied in the mentioned three-phased approach, and the feature extraction was accomplished through chromatic and Prewitt-edge highlights.

Download Full-text

Potential diagnosis of COVID-19 from chest X-ray and CT findings using semi-supervised learning

Physical and Engineering Sciences in Medicine ◽

10.1007/s13246-021-01075-2 ◽

2021 ◽

Author(s):

Pracheta Sahoo ◽

Indranil Roy ◽

Randeep Ahlawat ◽

Saquib Irtiza ◽

Latifur Khan

Keyword(s):

Supervised Learning ◽

X Ray ◽

Ct Findings ◽

Chest X Ray

Download Full-text

Multi-Feature Semi-Supervised Learning for COVID-19 Diagnosis from Chest X-Ray Images

10.1007/978-3-030-87589-3_16 ◽

2021 ◽

pp. 151-160

Author(s):

Xiao Qi ◽

David J. Foran ◽

John L. Nosher ◽

Ilker Hacihaliloglu

Keyword(s):

Supervised Learning ◽

X Ray ◽

Chest X Ray

Download Full-text

COVID-19 Chest X-Ray Detection Performance Through Variations of Wavelets Basis Function

Matrik Jurnal Manajemen Teknik Informatika dan Rekayasa Komputer ◽

10.30812/matrik.v21i1.1089 ◽

2021 ◽

Vol 21 (1) ◽

pp. 31-42

Author(s):

I Gusti Ayu Agung Diatri Indradewi ◽

Ni Wayan Sumartini Saraswati ◽

NI Wayan Wardani

Keyword(s):

Feature Extraction ◽

Error Rate ◽

Haar Wavelet ◽

Quality Parameters ◽

Classification Model ◽

Support Vector ◽

Daubechies Wavelet ◽

X Ray ◽

Chest X Ray ◽

Parameters Of Accuracy

Our previous work regarding the X-Ray detection of COVID-19 using Haar wavelet feature extraction and the Support Vector Machines (SVM) classification machine has shown that the combination of the two methods can detect COVID-19 well but then the question arises whether the Haar wavelet is the best wavelet method. So that in this study we conducted experiments on several wavelet methods such as biorthogonal, coiflet, Daubechies, haar, and symlets for chest X-Ray feature extraction with the same dataset. The results of the feature extraction are then classified using SVM and measure the quality of the classification model with parameters of accuracy, error rate, recall, specification, and precision. The results showed that the Daubechies wavelet gave the best performance for all classification quality parameters. The Daubechies wavelet transformation gave 95.47% accuracy, 4.53% error rate, 98.75% recall, 92.19% specificity, and 93.45% precision.

Download Full-text