A Filtering technique based on a DLMS algorithm for ultrasonography video

Abstract

It is well known that ultrasonography is a diagnostic method for visualizinginside human tissues by spreading ultrasounds and measuring their return time tothe sensor. However, the interface between the human skin and this ultrasoundtransducer attenuates the received signal and the medical image quality deterioratessignificantly. In this paper we propose a filtering technique in order to compensatethis attenuation. A finite impulse response filter (FIR) based on a Delayed LeastMean Square (DLMS) was optimized and implemented. The main contribution ofour work consists of finding the order and the coefficients of the filter that minimizethe attenuation error. We validate our method first on simulated data and later on areprogrammable FPGA device for a real time performance testing. Among others,we show that incrementing the order of the filter, not always is the best way toreduce image quality errors.