Micro-Displacement Detection using Echo-Signal Cepstrum Analysis for Medical Diagnosis

Abstract

The estimation of micro changes in morphological dimensions of thin biological structures (e.g. arterial wall, pulmonary pleura, sclera, etc) is a significant area of medical diagnosis where ultrasound has been applied more often each day; achieving higher resolutions on mechanical parameters which allow to provide information about biological tissue condition in early stage of some diseases. Digital processing techniques have been applied to evaluate their performance to detect these micro changes, some of them, with excellent results. In this paper, linear cepstrum technique was applied, as an alternative tool, to assess it performance in the estimation of micrometric separation between laminar structures. A basic mechanism controlled fixed increments (of 500 micrometers) of the separation between two low density polyethylene laminar layers that simulates biological membranes while an inspection system, based on ultrasound, was used to acquire the echo-signals. Previously to processing, the set of these signals were filtered and segmented, then they were added in pairs to be processed with cepstrum technique. In all cases, a quefrency peak value related to the inter-layers separation of 500 micrometers was found. Resultant distances were compared with reference measurements and an average error of 3.45 percent, i.e. 17.7 micrometers, was estimated.