M.H. Yap, E.A. Edirisinghe, and H.E. Bez (UK)
medical image analysis, ultrasound imaging, region-of-interest, FIR filter, thresholding segmentation.
This paper proposes a novel approach to initial lesion detection in ultrasound breast images. The objective is to automate the manual process of ROI labelling in Computer-aided Diagnosis (CAD). We propose the use of a Finite Impulse Response (FIR) filter and thresholding segmentation, in initial lesion detection and automated ROI labelling. A total of 360 ultrasound breast images have been used to evaluate performance of the proposed approach. Histogram equalization is used to pre-process the images followed by hybrid filtering and multifractal analysis. Subsequently a third order FIR filter is applied on the image proceeded by a thresholding segmentation stage. Finally the initial lesions are detected using a rule based approach. The accuracy of the automated Region of-Interest (ROI) labelling is measured by an overlap of 0.4 with the lesion outline compared to the lesions labelled by an expert radiologist. We compare the proposed method to two existing state-of-the-art method, namely the radial gradient index filtering technique, and the local mean technique. We conclude that the proposed method is more accurate and performs better than the selected benchmarks.
Important Links:
Go Back
