Arpita M Kadganchi, Geeta Hanji
Abstract: Ultrasound is widely used in medical applications. However, the problem in ultrasound images is they contain many speckles, and this make a sonographer hard to interpret the ultrasound image. During transmit & storage of Ultrasound images various other types of noise also get added up interfering with image analysis & interpretation. Our idea was to solve this problem by carrying out the pre-processing of images before proceed with further image processing techniques. In our work various image processing filters were used to remove the speckles, impulse & Gaussian noise so that the area of the region needed is clearer. The segmentation performed after pre-processing (identifying & removing the noise & enhancement) can help sonographer to analyze the ultrasound images qualitatively and quantitatively. In our work Linear and nonlinear filters were used for image pre-processing and their performance is compared with respect to the effect of segmentation of the ultrasound images. Few novel filters from both the categories were identified, implemented and feasible modifications were later incorporated. Then the segmentation process was carried out with the promising results were presented in the form of qualitative and quantitative dimensions in the final stage of our work.
Keywords: Ultrasound images, Filters, Thresholding process, Morphological process, Segmentation