FUSION OF BREAST IMAGES USING TEXTURAL INFORMATION AND OPTIMISED WAVELET TRANSFORMATION.

Image restoration is one of the steps in the image processing applications to restore the corrupted or noisy image into its original state. It plays an important role in all types of applications like segmentation, classification, etc. and in fields such as remote sensing, medical imaging. Restoration is an important process in medical imaging field to improve the quality of image for the diagnosis of disease. In this, a fusionbased approach is proposed for restoring and improving the quality of breast cancer images. The X-ray image of breast (mammogram) and magnetic resonance of breast image of same person is fused to improve the abnormality detection in the breast. The mammogram image is restored using top-hat filtering and Gabor filter to preserve the texture properties of the image. A contrast enhancement is performed to magnetic resonance breast image to hold the textural properties which is responsible for the abnormal detection in the breast image. The image fusion is performed to combine to two images into a single image which holds more information of both the images as compared to the single images. The restored mammogram and Magnetic Resonance Imaging (MRI) images are fused to improve the quality of image for the better cancer diagnosis in the breast images. The discrete wavelet transform is used for the fusion of restored images. The type of wavelet used for the fusion is determined with the help of Particle swarm optimisation (PSO) by maximising the entropy of fused image. By using PSO, the daubechies family db45 and symlet family sym1 is suitable for the image fusion process among the three wavelet families like daubechies (db1 to db45), Coiflet (coif1 to coif5) and symlet (sym1 to sym10). The proposed method is simulated using MATLAB 2018a and the performance is evaluated using entropy, mean, standard deviation and variance. The proposed method are able to achieve higher entropy of 8.1 for the fused method. Hence, the proposed fusion-based restoration method is able to improve the quality of image. [ABSTRACT FROM AUTHOR]