Experimental Analysis

Data Set Description

The verification of the projected GLCM-PSO-SVM method is carried out by applying standard open access data set BRATS 2015, which is comprised of a collection of MRI images [18]. The execution portion is implemented by the MATLAB tool with 1.70 GHz GPU Processor and 6 GB internal RAM. The applied data set is composed of three different brain MRI image sub data sets such as Training, Leader Board, and Challenge, as shown in Fig. 10.3.

Here, it is applied with the Training and Challenge data set. It is comprised of high-grade tumor (HGT) images and low-grade tumor (LGT) images with ground truth images from different radiologists. The MRI images of the training data set

Sample images

FIGURE 10.3 Sample images.

have been used in classification training of the proposed scheme. Therefore, access can be made via online to calculate the outcome of provided techniques. The data set has images of two types: benign and malignant.

Evaluation Metrics

To examine the outcome of the GLCM+PSO-SVM approach, a comparative investigation is processed among output images with ground truth images. According to the expert’s knowledge, the ground truth images are produced. An exactly classified nontumor image has been presented as true positive (TP) and an accurately classified tumor image is expressed as true negative (TN). Hence, the incorrectly classified tumor image is given as false positive (FP) and the incorrectly classified nontumor image is presented as false negative (FN). Such variables are computed according to the comparison with ground truth images.

A collection of values applied in estimating the function as given below ranges from 0 and 100. The values applied are

Results Analysis

Table I0.l provides a relative examination of projected and existing models on the classification of ВТ images. Fig. Ю.4 examines the results analysis of distinct techniques [19] in terms of sensitivity. The method developed by Anitha et al. (2017) demonstrated its ineffective outcome with a minimum sensitivity value of 91.20%. Next, the technique developed by Urban et al. (2014) exhibited a slightly higher sensitivity value of 92.60%. Afterward, the methods devised by Pereira et al. (2016) and

TABLE 10.1

Comparisons of Proposed with State of the Arts Methods









Selvapandian et al. (2018)




Anitha et al. (2017)




Pereira et al. (2016)




Urban et al. (2014)




Islam et al. (2013)




Sensitivity analysis of distinct models

FIGURE 10.4 Sensitivity analysis of distinct models.

Islam et al. (2013) have resulted in a moderate and near-identical sensitivity value of 94.20% and 94.30%, respectively. The model introduced by Selvapandian et al. (2018) has shown compete results by offering a near-optimal sensitivity value of 96.20%. However, the GLCM-PSO-SVM model has reached a maximum sensitivity of 91.20%.

Fig. 10.5 investigates the analysis of the results of distinct techniques with respect to specificity. The method coined by Urban et al. (2017) depicted its ineffective result with a minimum specificity value of 93.00%. Next, the technique developed by Anitha et al. (2017) exhibited a slightly higher specificity value of 93.40%. Later,

Accuracy analysis of distinct models

FIGURE 10.6 Accuracy analysis of distinct models.

the methods devised by Pereira et al. (2016) resulted in moderate and near-identical specificity values of 94.40%. The model developed by Selvapandian et al. (2018) and Islam et al. (2013) has shown compete results by offering a near optimal specificity value of 95.10%. Therefore, the GLCM-PSO-SVM systems has reached a maximum specificity of 96.12%.

Fig. 10.6 forecasts the results analysis of distinct techniques by means of accuracy. The technique deployed by Anitha et al. (2017) demonstrated its ineffective outcome with a minimum accuracy value of 93.30%. Next, the technique developed by Urban et al. (2014) exhibited a slightly higher accuracy value of 93.30%. Then, the methods devised by Pereira et al. (2016) and Islam et al. (2013) have resulted in moderate and near identical accuracy values of 94.60% and 95.90%, correspondingly. The model introduced by Selvapandian et al. (2018) has shown compete results by offering a near-optimal accuracy value of 96.40%. But, the GLCM-PSO-SVM model has reached a maximum accuracy of 97.96%.


This chapter has introduced a proficient GLCM-PSO-SVM model to identify and classify BTs with the use of IoHT and CC. The projected model operates on three stages: preprocessing, feature extraction, and classification. GLCM-based feature extraction and PSO-SVM-based image classification processes are carried out. The presented PSO-SVM model is validated using a set of images from the BRATS data set. The simulation outcome ensured that the PSO-SVM model is effective in terms of sensitivity, accuracy, and specificity. The experimental outcome ensured that the GLCM-PSO-SVM model has reached maximum classifier results with the highest sensitivity of 97.08%, a specificity of 96.12% and an accuracy of 97.96%. As part of future work, the presented technique results could be improvised by the use of segmentation approaches.


[1] Srivastava, G., Parizi. R. M„ and Dehghantanha. A., The future of blockchain technology in healthcare internet of things security, in Blockchain Cybersecurity, Trust and Privacy, Springer. Cham, 2020. 161-184.

[2] Farahani. B.. Firouzi, F„ and Chakrabarty, K.. Healthcare IOT. in Intelligent Internet of Things Springer, Cham, 2020, 515-545.

[3] Sahu, S. N.. Moharana. M.. Prusti, P. C.. Chakrabarty, S„ Khan. F., and Pattanayak, S. K., Real-time data analytics in healthcare using the Internet of Thing, in Real-Time Data Analytics for Large Scale Sensor Data, Academic Press. 2020, 37-50.

[4] Paul. P„ Dutta, N.. Biswas, B. A., Das. M., Biswas, S.. Khalid, Z., and Saha. H. N. An Internet of Things (IoT) Based System to Analyze Real-time Collapsing Probability of Structures, in 2018 IEEE 9th Annual Information Technology, Electronics and Mobile Communication Conference (IEMCON), IEEE, 2018, 1070-1075.

[5] Maitra, M.. and Chatterjee, A. A Slantlet transform based intelligent system for magnetic resonance brain image classification. Biomedical Signal Processing and Control, 1 (4). 299-306. 2006.

[6] Ma, Y.. Wang, Y.. Yang. J„ Miao. Y.. and Li, W. Big health application system based on health internet of things and big data, IEEE Access, 5, 7885-7897, 2017.

[7] Abbasi, S„ and Tajeripour. F. Detection of brain tumor in 3D MRI images using local binary patterns and histogram orientation gradient. Neurocomputing, 219, 526-535, 2017.

[8] El-Dahshan, E.S.A., Mohsen. H.M., Revett. K., and Salem, A.B.M. Computer-aided diagnosis of human brain tumor through MRI: A survey and a new algorithm, Expert Systems with Applications, 41 (11), 5526-5545, 2014.

[9] Menze. B.. Reyes, M., Jakab, A., Gerstner, E., Kirby, J., Kalpathy-Cramer, J., and Farahani. K.. NCI-MICCAI challenge on multimodal brain tumor segmentation, in Proceedings of NCI-MICCAI BRATS, Nagoya, Japan. 2013.

[10] Geremia, E.. Zikic. D„ Clatz. O., Menze, B. H„ Glocker. B.. Konukoglu. E„ Shotton.

J., Thomas, О. M., Price, S. J., Das, T., and Jena, R„ Classification forests for semantic segmentation of brain lesions in multi-channel MRI, in Decision Forests for Computer Vision and Medical Image Analysis, Springer, London, 2013, 245-260.

[11] Hussein, E„ and Mahmoud, D. Brain tumor detection using artificial neural networks. Journal of Science Technology, 13, 31-39, 2012. http://www.sustech.edu/staff publications/20130323072020220.pdf.

[12] Sharma, Y. and Chhabra. M. An Improved Automatic Brain Tumor Detection System. International Journal of Advanced Research in Computer Science and Software Engineering, 5 (4), 11-15, 2015.

[13] Shankar. K.. Elhoseny. M.. Lakshmanaprabu, S. K„ Ilayaraja, M„ Vidhyavathi, R. M.. Elsoud. M. A., and Alkhambashi. M. Optimal feature level fusion based ANFIS classifier for brain MRI image classification. Concurrency and Computation: Practice and Experience, 32 (1), 1-12, 2018.

[14] Elhoseny, M.. and Shankar. K. Optimal bilateral filter and convolutional neural network based denoising method of medical image measurements. Measurement, 143, 125-135, 2019.

[15] Shankar. K.. Lakshmanaprabu, S. K., Khanna, A.. Tanwar. S., Rodrigues, J. J.. and Roy, N. R. Alzheimer detection using Group Grey Wolf Optimization based features with convolutional classifier, Computers & Electrical Engineering, 77, 230-243, 2019.

[16] Elhoseny, M„ Shankar. K., and Uthayakumar, J. Intelligent diagnostic prediction and classification system for chronic kidney disease. Scientific Reports, 9 (1). 1-14, 2019.

[17] Lakshmanaprabu. S. K.. Mohanty, S. N.. Krishnamoorthy. S., Uthayakumar. J.. and Shankar, K. Online clinical decision support system using optimal deep neural networks. Applied Soft Computing, 81, 105487, 2019.

[18] MRBrainS18, Grand Challenge on MR Brain Segmentation at MICCAI, 2018, available at https://mrbrainsl8.isi.uu.nl/

[19] Selvapandian, A., and Manivannan, K. Fusion based glioma brain tumor detection and segmentation using ANFIS classification. Computer Methods and Programs in Biomedicine, 166, 33-38, 2018.

< Prev   CONTENTS   Source   Next >