Brain Tumor Classification Based on Hybrid Optimized Multi-features Analysis Using Magnetic Resonance Imaging Dataset

Figures & data

Table 1. Procedural Steps of the HBTC framework

Algorithm 1:Hybrid Brain Tumor Classification Procedure
Begin
Main {	step1: Input ɛ Brain Tumor MRI dataset step2: For 1 to Size-of Brain MRI dataset { Acquire brain tumor MRI Pre-process the input image Segment the tumor region Create segmented ROI Extract RLM, COM, Gradient features } End-For step3: Optimize via (MI + F + PA) + CFS step4: Multi-features analysis & classification step5: Evaluation
}End-Main

Figure 1. Sample Images of Four Brain Tumor Types.

Figure 2. Sample Effects of Image Preprocessing.

Table 2. TACS Model

Algorithm 2: threshold and clustering-based segmentation (TACS)
Input:	Brain Image Dataset
Parameter:	MRI Dataset {Brain Image Dataset}
Output:	Clustered Images (C_I) and Region of Interest (R_I)
1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18.	Function TACS (MRI Dataset) Begin for Total Pixel Measure of whole image (TPM) do B_P ← set threshold for back-ground Pixels P_G ← measure Pixel Gray Intensity T_C_C ← initialize Total_Cluster_Count with 0 if P_G > B_P then T_C_C ← increment Total_Cluster_Count F_G ← calculate fore-ground Pixels R_I ← Region of Interest from image else NOT count Segmentation Area B_P ← calculate Back-ground Pixel end if end for return (C_I, R_I) end Begin end Function

Figure 3. Overall Model of TACS Methodology with taking ROI.

Table 3. F+ MI+PA+CFS based selected features

Sr. No.	Feature
1.	S(1_0)SumAverg
2.	S(5_5)Correlat
3.	S(1_-1)SumAverg
4.	S(4_0)SumAverg
5.	S(0_5)SumAverg
6.	S(0_4)SumAverg
7.	S(0_3)Correlat
8.	S(0_2)Correlat
9.	S(0_1)InvDfMom

Figure 4. The Complete HBTC Framework.

Table 4. Parameters values of MLP

MLP Parameters	Values of Parameters
Hidden-Layers	1
Input Layers	15
Neurons	10
Momentum	0.2
Learning Rate	0.3
Threshold Validation	20
Epochs	500

Table 5. HBTC Machine Vision Classifiers based on 10 × 10 ROI

Classifiers	Kappa_ Statistics	T_P_Rate	F_P Rate	R_C	T_N_I	Time. (Sec)	Overall Accuracy
MLP	0.5307	0.648	0.117	0.827	1000	0.74	64.8%
J48	0.534	0.643	0.119	0.790	1000	0.02	64.3%
MB	0.5227	0.642	0.119	0.820	1000	0.13	64.2%
RT	0.496	0.622	0.126	0.708	1000	0	62.2%

Table 6. Confusion Matrix for MLP based on 10 × 10 ROI

Classes	cyst	Glioma	menin	meta	Total
cyst	240	0	2	8	250
glioma	2	200	30	18	250
menin	1	37	193	19	250
meta	39	78	15	118	250

Table 7. HBTC Machine Vision Classifiers based on 15 × 15 ROI

Classifiers	Kappa_ Statistics	T_P_Rate	F_P Rate	R_C	T_N_I	Time. (Sec)	Overall Accuracy
MLP	0.852	0.889	0.037	0.975	1000	0.75	88.9%
J48	0.86	0.895	0.035	0.956	1000	0.02	89.5%
MB	0.8373	0.878	0.041	0.971	1000	0.05	87.8%
RT	0.752	0.814	0.062	0.876	1000	0	81.4%

Table 8. Confusion Matrix for J48 based on 15 × 15 ROI

Classes	cyst	glioma	menin	meta	Total
cyst	239	5	3	3	250
glioma	5	245	0	0	250
menin	3	77	167	3	250
meta	4	0	2	244	250

Table 9. HBTC Machine Vision Classifiers based on 20 × 20 ROI

Classifiers	Kappa_ Statistics	T_P_Rate	F_P Rate	R_C	T_N_I	Time. (Sec)	Overall Accuracy
MLP	0.9773	0.983	0.006	0.998	1000	1.19	98.3%
J48	0.9573	0.968	0.011	0.981	1000	0.01	96.8%
MB	0.944	0.958	0.014	0.971	1000	0.05	95.8%
RT	0.9307	0.948	0.017	0.965	1000	0	94.8%

Table 10. Confusion Matrix for MLP based on 20 × 20 ROI

Classes	cyst	glioma	menin	meta	Total
cyst	246	2	1	1	250
glioma	1	249	0	0	250
menin	5	0	241	4	250
meta	2	1	0	247	250

Figure 5. Comparison Graph of MLP on ROIs of sizes 10 × 10, 15 × 15 and 20 × 20.

Figure 6. Comparison Graph of Overall Classicification Performance of Four Classifers.

Table 11. Comparison of HBTC framework with other streamed classification techniques

Source Citation	Key Features Extracted	Classification	Overall Acc
(Mudda, Manjunath, and Krishnamurthy 2020a)	MRI brain, GLRLM, CSLBP,	ANN	94%
(Seere and Karibasappa 2020)	MRI brain, Homogeneity, Shape, Color, Size, Energy, Contrast	SVM	85.32%
(Ortiz-Ramón et al. 2020)	MRI brain, GLCM, GLRLM, GSSZM NGTDM, LBP.	Linear SVM	89.6%
(Gupta and Sasidhar 2020)	MRI brain, SFTA based texture features	SVM	87%
(Zacharaki et al. 2009)	MRI brain, Statistical & Shape Based, Intensity based, Gabor.	KNN, SVM	90.6%
(Ullah, Batool, and Gilanie 2018)	RD-BVH Brain MRI, Intensity, shape, Texture Features	SVM	97%
(Selvaraj et al. 2007)	MRI brain, Statistical, GLCM	SVM, MPL,	77 to 98%
(Gilanie et al. 2019)	RD-BVH Brain MRI, Texture Features	ANN	86% to 92%
(Tiwari et al. 2017)	MRI brain, Textural Features, GLCM, GLRLM, Gabor, LTEM, FOS	ANN	92.43%
(George and Manuel 2019)	MRI brain, Histogram, First order statistics, GLCM	DNN	91.92%
(Mohsen et al. 2018)	MRI brain, DWT, PCA,	DNN	96.97%
(Bhanumathi and Sangeetha 2019)	MRI brain, Shape, Intensity, Color, Statistics	CNN	97.10%
(Gumaei et al. 2019)	MRI brain, Gabor, GWT, GIST-descriptor, N-GIST (normalized)	RELM	94.233%
Proposed Methodology	HBTC framework	MLP	98.3%

Mudda, M., R. Manjunath, and N. Krishnamurthy. 2020a. Brain Tumor Classification Using Enhanced Statistical Texture Features. IETE Journal of Research 1–12. doi:10.1080/03772063.2020.1775501.

 Web of Science ®Google Scholar

Seere, S. K., and K. Karibasappa. 2020. Threshold Segmentation and Watershed Segmentation Algorithm for Brain Tumor Detection Using Support Vector Machine. European Journal of Engineering Research and Science 5 (4):516–19. doi:10.24018/ejers.2020.5.4.1902.

 Google Scholar

Ortiz-Ramón, R., S. Ruiz-España, E. Mollá-Olmos, and D. Moratal. 2020. Glioblastomas and Brain Metastases Differentiation Following an MRI Texture Analysis-Based Radiomics Approach. Physica Medica 76:44–54. doi:10.1016/j.ejmp.2020.06.016.

 PubMed Web of Science ®Google Scholar

Gupta, M., and K. Sasidhar. 2020. “Non-Invasive Brain Tumor Detection Using Magnetic Resonance Imaging Based Fractal Texture Features and Shape Measures.” Proceedings of 3rd International Conference on Emerging Technologies in Computer Engineering: Machine Learning and Internet of Things, ICETCE 2020 Jaipur, India (IEEE), 93–97. doi: 10.1109/ICETCE48199.2020.9091756.

 Google Scholar

Zacharaki, E. I., S. Wang, S. Chawla, D. Soo Yoo, R. Wolf, E. R. Melhem, and C. Davatzikos. 2009. Classification of Brain Tumor Type and Grade Using MRI Texture and Shape in a Machine Learning Scheme. Magnetic Resonance in Medicine 62 (6):1609–18. doi:10.1002/mrm.22147.

 PubMed Web of Science ®Google Scholar

Ullah, H., A. Batool, and G. Gilanie. 2018. Classification of Brain Tumor with Statistical Analysis of Texture Parameter Using a Data Mining Technique. International Journal of Industrial Biotechnology and Biomaterials 4 (2): 22–36 . February 2019.

 Google Scholar

Selvaraj, H., S. Thamarai Selvi, D. Selvathi, and L. Gewali. 2007. Brain Mri Slices Classification Using Least Squares Support Vector Machine. IC-MED International Journal of Intelligent Computing in Medical Sciences and Image Processing 1 (1):21–33. doi:10.1080/1931308X.2007.10644134.

 Google Scholar

Gilanie, G., U. I. Ijaz Bajwa, M. M. Mahmood Waraich, and Z. Habib. 2019. Automated and Reliable Brain Radiology with Texture Analysis of Magnetic Resonance Imaging and Cross Datasets Validation. International Journal of Imaging Systems and Technology 29 (4):531–38. doi:10.1002/ima.22333.

 Web of Science ®Google Scholar

Tiwari, P., J. Sachdeva, C. K. Kamal Ahuja, and N. Khandelwal. 2017. Computer Aided Diagnosis System-A Decision Support System for Clinical Diagnosis of Brain Tumours. International Journal of Computational Intelligence Systems 10 (1):104–19. doi:10.2991/ijcis.2017.10.1.8.

 Web of Science ®Google Scholar

George, N., and M. Manuel. 2019. “A Four Grade Brain Tumor Classification System Using Deep Neural Network.” In 2019 2nd International Conference on Signal Processing and Communication (ICSPC), Coimbatore, INDIA (IEEE), 127–32. doi: 10.1109/ICSPC46172.2019.8976495.

 Google Scholar

Mohsen, H., E. S. A. El-Dahshan, E. S. M. El-Horbaty, and A. B. M. Salem. 2018. Classification Using Deep Learning Neural Networks for Brain Tumors. Future Computing and Informatics Journal 3 (1):68–71. doi:10.1016/j.fcij.2017.12.001.

 Google Scholar

Bhanumathi, V., and R. Sangeetha. 2019. “CNN Based Training and Classification of MRI Brain Images.” In 2019 5th International Conference on Advanced Computing and Communication Systems, ICACCS 2019 15-16 March 2019 (IEEE) Coimbatore, India, 129–33. doi: 10.1109/ICACCS.2019.8728447.

 Google Scholar

Gumaei, A., M. M. Mehedi Hassan, M. R. Rafiul Hassan, A. Alelaiwi, and G. Fortino. 2019. A Hybrid Feature Extraction Method with Regularized Extreme Learning Machine for Brain Tumor Classification. IEEE Access 7:36266–73. doi:10.1109/ACCESS.2019.2904145.

 Web of Science ®Google Scholar