Abstract:
In this project, we have studied the use of wavelet transform (WT) for processing audio, image and MRI signal. For audio signal we used different wavelets such as Haar, Daubechies, Biorthgonal, Symmlet and Coiflet for decomposing and reconstructing selected signal. We measured the effect of several wavelets for evaluating the quality of the reconstructed signal. The parameters of evaluations are Peak Signal to Noise Ratio (PSNR), Normalized Root Mean Square Error (NRMSE), Signal to Noise Ratio (SNR), Retained energy (RE). We observed that Haar gives highest value of PSNR but NRMSE, SNR is low. For coiflet gave lowest value of NRMSE, SNR but PSNR is highest as compare to haar . Therefore haar is best for our audio data. For image, we used Biorthogonal 4.4 wavelets and different type of compression methods such as Embedded Zero tree Wavelet (EZW), Set Partitioning in Hierarchical Trees (SPIHT) Coding, Wavelet Difference Reduction (WDR), Adaptively scanned (ASWDR), Spatial-orientation Tree Wavelet (STW), Set Partitioning in Hierarchical Trees-3D (SPIHT-3D) Coding, and Wavelet level metropolis Monte Carlo (LVL-MMC) etc. We measured the effect of several compression methods for evaluating the quality of the reconstructed image. The parameters evaluations are Mean square Error (M.S.E), are Peak Signal to Noise Ratio (PSNR),Max Error, and Bit per pixel (BPP),Compression ratio. We observed that EZW gave lowest value of M.S.E, Max. Error but PSNR, for STW gave almost similar results like EZW. So, EZW method is best for our image. For MRI signal, we used different wavelets such as Haar, Deubachies, Biorthgonal, Symmlet and Coiflet for decomposing and reconstructing signal. We measured the effect of several wavelets for evaluating the quality of the reconstructed signal. We calculated the error between the original and the reconstructed image. We observed that Haar gives lowest value of Error where PSNR is high. Coiflet gave largest value of Error as compare to haar. So, Haar is best for our MRI signal.