Contour Detection for the Breast Tumor in Ultrasonic Images Using Watershed Segmentation

Abstract

Due to the ultrasonic examination would not cause any side effect upon human’s body. Moreover, because the low price, convenience, prevalence, and timeliness of ultrasonic scanner, the ultrasonic instruments grow into essential for hospitals. The ultrasound became the most acceptable procedure for patients in the different types of digital medical image. The ultrasonic image is also an efficient instrument of the clinical physicians to diagnose the nidus at an earlier stage. Automatic contour finding for the breast tumors in the ultrasonic images may assist physicians without experience in making a correct diagnosis. Unfortunately, the digital ultrasonic image always comprises speckles, noise, and tissue-related textures, most traditional segmentation techniques for the ultrasonic images do not perform well. In this paper, we combined texture analysis techniques and the watershed segmentation to detect the contour of breast tumors in the ultrasonic images. The auto-correlation coefficients are applied as texture features to classify the breast ultrasound images by utilizing the self-organizing map (SOM). After the SOM model classifying the texture features, the watershed transform is used to detect the tumor contour. Computer simulation results show that the proposed method always found the similar contour with the manual sketch of the breast tumor in the ultrasonic images.