A Noninvasive System for Gliomas Diagnosis Based on Tamura Texture, Discrete Wavelet Transformation and Pyramid Histogram of Oriented Gradient

The glioma is regarded as one of the most common malignant brain tumors, which is a serious threat to the life of patients. The early detection of gliomas can contribute to make up a suitable surgery scheme and thus improve the survival rate of patients. Conventional methods to diagnosis gliomas rely mostly on the clinical experiences of radiologists, which is of low efficiency and accuracy. To improve the detection accuracy of the gliomas, an automatic diagnosis system based on T2-weighted brain images is presented in this paper. In this system, brain images are labeled with normal, glioma and the other kinds of tumors, in addition, hybrid features including discrete wavelet transformation (DWT), Tamura texture and pyramid histogram of oriented gradient (PHOG) are extracted, and ant colony algorithm (ACA) is combined with support vector machine (SVM) to build up classifiers. The experiment results show that the accuracy, specificity and sensitivity of this method can reach 91.11%, 86.91% and 94.21%, respectively.

[1]  K. Satya Prasad,et al.  Advanced Morphological Technique for Automatic Brain Tumor Detection and Evaluation of Statistical Parameters , 2016 .

[2]  Mehdi Jafari,et al.  Automatic Tumor Classification in Brain MRI Images Using Genetic Algorithm and Artificial Neural Network , 2014 .

[3]  Pritee Khanna,et al.  A non-invasive and adaptive CAD system to detect brain tumor from T2-weighted MRIs using customized Otsu's thresholding with prominent features and supervised learning , 2017, Signal Process. Image Commun..

[4]  Sharda Vashisth,et al.  Brain Tumor Detection and Classification with Feed Forward Back-Prop Neural Network , 2017, ArXiv.

[5]  Muhammad Sharif,et al.  A distinctive approach in brain tumor detection and classification using MRI , 2017, Pattern Recognit. Lett..

[6]  Kalpana U. Rathod,et al.  Automated Brain Tumor Detection and Brain MRI Classification Using Artificial Neural Network-A Review , 2016 .

[8]  R. Chang,et al.  Support vector machines for diagnosis of breast tumors on US images. , 2003, Academic radiology.

[9]  Yudong Zhang,et al.  AN MR BRAIN IMAGES CLASSIFIER VIA PRINCIPAL COMPONENT ANALYSIS AND KERNEL SUPPORT , 2012 .

[10]  Zhihao Shao,et al.  Adaptive Harris corner detection algorithm based on iterative threshold , 2017 .

[11]  Abdellatif Mtibaa,et al.  Modified support vector machines for MR brain images recognition , 2013, 2013 International Conference on Control, Decision and Information Technologies (CoDIT).

[12]  Takahiro Hara,et al.  A Multi-Objective Optimization Scheduling Method Based on the Ant Colony Algorithm in Cloud Computing , 2015, IEEE Access.

[13]  Ahmed Bouridane,et al.  Multispectral palmprint recognition using Pascal coefficients-based LBP and PHOG descriptors with random sampling , 2017, Neural Computing and Applications.

[14]  Vinod Kumar,et al.  Segmentation, Feature Extraction, and Multiclass Brain Tumor Classification , 2013, Journal of Digital Imaging.