27 August 2024

Segmentation of an Ultrasonic Reflector Image Based on Analyzing the Proximity of Its Histogram to the Rayleigh Distribution

Currently, reflector recognition (classification) systems based on artificial neural networks are being actively developed in order to increase the speed of preparation of the ultrasound testing protocol and reduce the influence of the human factor. For more efficient operation of neural networks, it is advisable to process images of reflectors in order to increase the signal-to-noise ratio of the image and its segmentation (clustering). The article proposes a segmentation method based on the construction of a binary mask that hides the reflector flare. The mask is created according to the condition of obtaining the image histogram with an appearance closest to the Rayleigh distribution. A genetic algorithm is used to solve the problem of finding the minimum. The model experiments show the effectiveness of this approach to segmentation of reflector images reconstructed from echo signals measured using antenna arrays. A method based on analyzing the flare amplitudes of images reconstructed using various acoustic schemes is used to determine the reflector type.

Read more >> (файл PDF, 2 Мб) 

Date: 27 August 2024
File: pdf / 1.87 MB
Language: English
Segmentation of an Ultrasonic Reflector Image Based on Analyzing the Proximity of Its Histogram to the Rayleigh Distribution DOWNLOAD
Feedback Form