Reduced Speckle noise on Medical Ultrasound Images Using Cellular Neural Network

Abstract

Speckle noise is indispensable get from ultrasound image. In general tends to reduce the image resolution and contrast. In addition to the doctor diagnosis, is lacking for judgment accuracy. This paper is reduced the speckle noise and enhanced boundary of a tumor in the medical ultrasound images. The proposed method is valuated using numerical phantom simulating ultrasound B-mode images, and the effect is confirmed by applying to medical ultrasound images. Therefore, some important features such as tissue boundaries and small tumors may be overlooked. A cellular neural network which is a kind of recurrent neural network can deal with images by the weight of neurons called a cell. It could be obtained more detail images recognition compared with the previous studies. Determination template parameters of the cellular neural network for ultrasound image processing are discussed. The experimental results show effectiveness of applying the proposed method to boundary enhancement and the speckle noise reduction of medical ultrasound image.