Bonfring International Journal of Advances in Image Processing
Online ISSN: 2277-503X | Print ISSN: 2250-1053 | Frequency: 4 Issues/Year
Impact Factor: 0.245 | International Scientific Indexing(ISI) calculate based on International Citation Report(ICR)
Wavelet Packet Based MRI Brain Tumor Discrimination and Classification
M. Anantha Valli, Allin Christe and Dr.A. Kandaswami
Abstract:
In this paper, classification of MRI brain tumor images based on texture features obtained from the two-level wavelet packet decomposition is proposed. Brain tumor a serious brain disorder is to be properly detected before for any further treatment. Normally a diseased region has chaotic or rougher structure than normal region. Image texture offer better information on the health of the examined region by measuring its features. So here the texture of the MRI image is analyzed using wavelet transform and wavelet packet decomposition and the results are compared. Ability of wavelets to discriminante different frequencies and to preserve signal details at different resolutions is the main reason for using wavelet transform and wavelet packet decomposition. Since the most significant information of a texture often appears in the high frequency channels, these high frequency channels are taken into account. Also wavelet and wavelet packet subbands have the ability to discriminate different texture patterns from the extracted features. Various first order statistical features such as mean, variance, entropy and energy are extracted from the decomposed sub-bands and are analyzed for discrimination among the tumors. Wavelet transform classifies the primary benign tumor with an accuracy of 83.33%, primary malignant type with 86.67% and secondary with 84.61% accuracies. Using wavelet packet decomposition the descendents of LH band that provides better discrimination accuracies to 87.5%, 90%, and 88.46% for primary benign, primary malignant and secondary type tumors respectively.
Keywords: Wavelet Transform, Wavelet Packet Decomposition, Texture Analysis, Feature Extraction
Volume: 2 | Issue: Special Issue on Communication Technology Interventions for Rural and Social Development
Pages: 72-79
Issue Date: February , 2012
|