Automated classification of types of brain tumor in T1-weighted MR images: a thorough comparative study
Undoubtedly, early detection and characterization of brain tumor is critical in clinical practices. Automated diagnosis using neuroimaging tool like MRI guided by machine learning approaches has been the focus of numerous researches. In this study, various feature extraction, dimensionality reductio...
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Conference or Workshop Item |
| Language: | English |
| Published: |
2020
|
| Subjects: | |
| Online Access: | http://eprints.utm.my/id/eprint/89883/1/LimJiaQi2020_AutomatedClassificationofTypesofBrainTumor.pdf http://eprints.utm.my/id/eprint/89883/ http://dx.doi.org/10.1063/5.0018056 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Undoubtedly, early detection and characterization of brain tumor is critical in clinical practices. Automated diagnosis using neuroimaging tool like MRI guided by machine learning approaches has been the focus of numerous researches. In this study, various feature extraction, dimensionality reduction and supervised classification models are explored, evaluated and compared under different finite number of features to identify the optimal pathway/pipeline for classification of types of brain tumor, namely meningioma, glioma and pituitary tumor. The performance metrics utilized include accuracy, Kappa statistic, sensitivity, precision, F-measure, training time and test time. Results show that RBF SVM (pairwise coupling) under 80 PLS features achieved the highest average accuracy (95.02% ± 0.19%) among all other machine learning pipelines. |
|---|