Into the ultrasound deconvolution using CLEAN algorithm - statistical analysis of scatters detection

The quality of the reproduced tissues in the ultrasound medical imaging is reduced by the imperfections in tissues and the finite frequency bandwidth and pulse width of the ultrasound transducer. Constrained by these constructive limitations, for improvement of details in the final image, the post processing methods are usually applied. In this paper we present a two step 1D blind deconvolution method based on the well known time domain CLEAN algorithm. In the first step the Point Spread Function is estimated and secondly, we use this estimation to extract the reflectivity function of the tissues. The simulations was oriented to evaluate statistically the method according to the percentage of real detection of scatters, the number of false alarms - which represent the quantity of the artefacts in the final result and the means square error. Also, the algorithm is compared with the classical Tikhonov regularization in term of above presented criteria and mean square error (MSE) - a global estimator between reflectors position and their amplitude.