International Journal of Engineering Research ISSN:2319-6890)(online),2347-5013(print) Volume No.5, Issue No.7, pp : 586-588 1 July 2016 IJER@2016 doi : 10.17950/ijer/v5s7/708 Page 586 Removal of Speckle Noise Using Hybrid Filter Technique R. Amuda #1 , K. Bhuvaneshwari *2 Department of Mathematics, Mother Therasa Women’s University ,Kodaikanal – 624 102, Tamil Nadu, India. 1 amuda@jeppiaarinstitute.org Abstract: Filters provide an aid to visual interpretation of images. So the design of filters in image processing is important to de-noise the images, to emphasize edges and to transform the correct images Filters depend on the type of noise present on the images. A new hybrid Filter DWT filters followed by Average filter is introduced to de-noise the speckle noise and investigated the performance on the images . Keywords: DWT filters, Average Filters and cascaded filters. I. INTRODUTION Digital topology is a mathematical tool in image processing. It digitalise the continuous image to discrete which is acceptable by computers vision and computer graphics. Most images are affected by noise that is unexplained variation in data and disturbance in image intensity. Image analysis is simplified if this noises are filtered. The design of filters are framed using mathematical analysis sources of noise in the digital image [1]are  environmental conditions  Insufficient Light levels and sensor temperature  Interference in the transmission channel  If dust particles are present on the scanner screen Image noise can be classified[2] as  Impulse noise (Salt-and-pepper noise)  Amplifier noise (Gaussian noise)  Shot noise, Quantization noise (uniform noise)  Shot noise, Quantization noise (uniform noise),  Film grain, on-isotropic noise, Multiplicative noise (Speckle noise)  Periodic noise. De-noising filters can be categorized[4] in the following categories:  Averaging filter  Order Statistics filter  Adaptive filter The best filters are one which remove the noise completely from the image, while preserving the details. To develop filters which can preserve the features of the area of interest are in need . In this paper a hybrid filtering technique is adapted to remove speckle noise in images. Several different methods are used to eliminate speckle noise, based upon different mathematical models. but a new hybrid Filter DWT filters followed by Average filter is introduced to de-noise the speckle noise and investigated the performance on two different images . II. DISCRETE WAVELET TRNSFORM Different types of Wavelet transforms are existed for specific purposes.  Continuous wavelet transform  Discrete wavelet transform  Complex wavelet transform  Dual wavelet transform A discrete wave let transform(DWT) is a wavelet transform for which the wavelets are discretely sampled. It captures both frequency and location information. DWT transforms the changes in time extension, but not shape. There are different type of discrete Wavelet transforms.  Haar Wavelet transforms  Daubechies Wavelet transforms  The dual-tree complex Wavelets transforms Haar Wavelet transforms(HWT) For an input of 2 n numbers Haar Wavelet transforms, storing the difference and passing the sum. This processes is repeated recursively till to lead 2 n-1 difference and a final sum. This is the Fast wavelet transform, an alternate to Fast Fourier transform. Daubechies Wavelets Transforms (DWT) The formulation is based on the recurrence relations to generate progressively finer discrete samplings of an implicit mother wavelet function. Each resolution is twice that of previous scale. The Dual-Tree complex wave Transform(DCT) It is invariant and directionally selective in two and higher dimensions. Mean Filter It is a linear filter . it computes the mean value of the corrupted image. Then the center pixel intensity value is replaced by the mean value. This process is applied to all pixel values in the image.Fig-1b) shows the effect of using mean filter on speckle noise. Speckle noise Speckle noise is caused by signals from elementary scatters in medical literature, it is referred as ‘texture’ and it may contain useful diagnostic information. The filters filter the more relevant details of the images. So it is essential to develop filters which can preserve the essential details of the images. Based on different mathematical models several hybrid filters are used. In this work a hybrid filtering technique is adapted to remove speckle noise. A cascaded filter technique is used . .cascade filter is a low pass filter and a high pass filter to produce a band pass filter. In this paper DWT filter followed by mean filter is introduced as a cascaded filter. this technique