International Research Journal of Engineering and Technology (IRJET) Volume: 10 Issue: 05 | May 2023
www.irjet.net
e-ISSN: 2395-0056 p-ISSN: 2395-0072
A Hybrid Approach for Morphological Study of Brain Tumors in MRI Images Rajan Chaturvedi1, Dr.R.K.Tiwari2 1Research scholar Department of Physics and Electronics, Dr. Rammanohar Lohia Avadh University, Ayodhya, Uttar
Pradesh, India,
2Corresponding Author: Professor Department of Physics and Electronics, Dr. Rammanohar Lohia Avadh University,
Ayodhya, Uttar Pradesh, India, --------------------------------------------------------------------------***------------------------------------------------------------------------aided diagnosis of brain tumors and abnormalities. The Abstract-Brain Tumor is one of the most deadly diseases
A Brain Tumor is a bunch or tassel of unusual grew cells inside the brain. These are of two types Cancerous and NonCancerous. Diagnosis of brain tumors in primary stages is crucial for the treatment and wellness of patients, Magnetic Resonance imaging (MRT) is the most popular and efficient imaging technology for the diagnosis of brain tumor. In MRI films various part of the brain exists with proximity with low contrast in a noisy environment. In such low contrast and noisy environment accurate diagnosis of shape, size, boundary, volume, and density of brain tumor is very challenging. Medical image analysis deals with the computer-
Segmentation and extraction of tumor area can be improved by fusion of morphological operator and watershed method. [1]. Tumor cells successfully separated from normal cells by using thresholding, watershed techniques which are generally used but the use of morphological operator provide better detection of tumor [2]. A segmentation accuracy of 82% to 97% and SNR volume from 20 to 44 can be achieved by soft thresholding DWT for enhancement and genetic algorithms for image segmentation [3]. A markerbased watershed segmentation produces a possible enhancement of extraction of brain tumor with high accuracy in MRI. Watershed algorithm for extraction of brain tumor has been identified as a more suitable method compared to existing methods like a combination of K-means and Fuzzy means, SVM and CRF, Kernel feature selection, etc. [4] A hybrid approach combining discrete wavelet transform, PCA for diminishing feature and SVM for classification of brain tumor reduces manual labeling time and avoid a human error with desirable accuracy. [5] Mathematical morphological reconstruction (MMR) based computer-aided detection of brain tumors shows good and accurate results [6]. A practical approach of the hybrid skull stripping method contains texture feature analysis, Fuzzy positivistic C-means (FPC), and morphological operator utilized for detection and extraction of brain tumor. And results have a high degree of accuracy, which is evaluated on the internet brain segmentation repository dataset. [7] The classification accuracy of 95% achieved for brain tumors into meningioma and glioma by using skull stripping with morphology and thresholding, segmentation performed with wavelet transform feature, self-organizing map, and watershed algorithm. Feature extraction is done by GLCM and classification by forwarding neural network. [8]A hybrid algorithm using neutrosophy and convolutional neural network used for classification of a brain tumor as benign and malignant. The finding of experimental results demonstrates that order classification with different classifiers and performs better with SVM having an average success of 95.62%. [9] The brain tumor detection in MR images performed by using super-resolution Fuzzy C-means
© 2023, IRJET
ISO 9001:2008 Certified Journal
and requires early detection for treatment. MRI & CT. Scan images are mostly used for the diagnosis of brain tumors. Medical image processing finds a great application for the auto-detection of a brain tumor in MR images. There are many existing techniques and algorithms for the detection and feature extraction of a brain tumor in MRI images. Morphological operators are a powerful tool for morphological study of objects like brain tumor in MR images and pseudo coloring is excellent tool for segmentation. The expansion region and boundary of the tumor can be identified accurately with the morphological operator. Pseudo coloring is a powerful art of state for segmentation of the objects in a crowded environment. In this paper a hybrid approach combining features of mathematical morphology and pseudo coloring used for the segmentation and boundary detection of a brain tumor in MRI. The experimental results of the used method evaluated against finding Neuroscience experts of King George Medical College Lucknow utter Pradesh. The proposed method exhibits a high degree of accuracy in tumor segmentation and boundary detection and experimental outcomes well agree with the finding of Neuroscience experts.
Keywords— Medical image processing, mathematical morphology, Erosion, Dilation, pseudo coloring.
I.
Introduction
|
Impact Factor value: 8.226
|
|
Page 1672