A novel approach to suppression of ultrasonic speckle based on a combi
nation of segmentation and optimum L-filtering is presented. With the
aid of a suitable modification of the learning vector quantizer (LVQ)
neural network, the image is segmented in regions of(approximately) ho
mogeneous statistics. For each of the regions a minimum mean-squared-e
rror (MMSE) L-filter is designed, by using the histogram of grey level
s as an estimate of the parent distribution of the noisy observations
and a suitable estimate of the (assumed constant) original signal in t
he corresponding region, Thus, a bank of L-filters results; with each
of them corresponding to and operating on a different image region. Si
mulation results from both simulated and real B-mode ultrasonic images
are presented, which verify the (qualitative and quantitative) superi
ority of our technique over a number of commonly used speckle filters.