Deep Learning Based Image Classification and Abnormalities Analysis of MRI Brain Images
| Autor: | Muthu Krishnammal P, S. Selvakumar Raja |
|---|---|
| Rok vydání: | 2019 |
| Předmět: |
Contextual image classification
business.industry Computer science Deep learning Feature extraction ComputingMethodologies_IMAGEPROCESSINGANDCOMPUTERVISION k-means clustering Wavelet transform Image processing Pattern recognition Convolutional neural network ComputingMethodologies_PATTERNRECOGNITION Segmentation Artificial intelligence business |
| Zdroj: | 2019 TEQIP III Sponsored International Conference on Microwave Integrated Circuits, Photonics and Wireless Networks (IMICPW). |
| DOI: | 10.1109/imicpw.2019.8933239 |
| Popis: | Intensive researches are being carried out to study the abnormalities present in the brain structures and to detect the type of tumors based on the statistical and textural features extracted from the medical images. In MRI imaging, the images may be clear but the clinicians have to quantify the size and location of the tumors for further treatment planning. The image processing methodologies and the deep learning methods aid the different stages of treatment such as pre and postsurgical procedures. The images captured by Magnetic Resonance Imaging systems are processed by the different software based algorithms in order to segregate the malicious tumor regions from the non-tumor regions. The proposed method includes three phases: feature extraction, image classification and segmentation of tumors. The important features of the images are extracted with Wavelet transform, a multi-resolution technique. The Convolutional Neural Network, which is a very popular deep learning method is used in image classification stage which is helpful in disease or lesion detection, image classification and then the segmentation methods are used to segregate the infected regions from the rest. The proposed method provides better accuracy in classification and segmentation stages. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|