| Autor: |
Sandhya, M., Balachander, Bhuvaneswari |
| Předmět: |
|
| Zdroj: |
AIP Conference Proceedings; 2024, Vol. 2871 Issue 1, p1-12, 12p |
| Abstrakt: |
Estimating textural distortion of pulmonary CT scans caused by COVID-19 occurrences is the primary goal of this investigation. The images from the dataset are acquired from Github.org. In all, 176 participants were surveyed for this study both non-COVID and COVID. In this analysis, the comparison between difference theoretic texture features and Grayscale co- occurrence medium to mine the changes in the texture of CT lung scans. The features are subjected to classification using K-Nearest Neighbour and Neural Network classifiers. In Differential Theoretic Texture Feature 10-fold cross validation is executed during the evaluation of classifiers acquired Area Under Curve ('AUC') (.82%), F1 Score (.73%), Accuracy (.73%), 'Recall' (.72%) and Classification (.72%) were obtain using KNN classifier 10-fold cross validation is performed during the evaluation of classifiers. Feature extraction techniques are applied on each of the lung CT images through MATLAB software. From the results, it is experiential that in normal topics surface bend is less owing to smooth surface and in COVID subjects texture deformation is highly outstanding to hankie defeat. The feature values obtained using DTTF are comparatively more significant than GLCM. The DTTF features such as absdiff has less P-value (0.007132) which is (p<0.05) compared to GLCM values. DTTF acquired area under curve In this study it is found that the Difference theoretic texture feature performs better in comparison with the Grayscale co-occurrence medium quality feature for the detection of COVID-19. [ABSTRACT FROM AUTHOR] |
| Databáze: |
Complementary Index |
| Externí odkaz: |
|
