An X-ray photon counting detector consistency calibration method based on deep learning

Abstract

The invention discloses an X-ray photon counting detector consistency calibration method based on deep learning. The method comprises the following steps: 1, bad pixel positioning: analyzing a projection image by using a clustering algorithm to obtain bad pixel coordinates; and 2, classifying and compensating the bad pixels: classifying the bad pixels by using correlation analysis to obtain the positions of the bad pixels of the detector, and calibrating the positions. And 3, noise elimination of the projection image: using the label data to train the convolutional neural network to eliminatenoise in the projection image, and completing consistency calibration of the X-ray photon counting detector. According to the method, the coordinates of the defective pixels of the detector are obtained by analyzing the rear-end projection image and are calibrated, so that the positions of the defective pixels of the detector can be more accurately determined, the quantum noise in the projection image is completely eliminated, and the method is more convenient and quicker than the existing front-end calibration.

Description

technical field [0001] The invention belongs to the calibration technology of detectors, in particular to the consistency calibration technology of X-ray photon counting detectors. Background technique [0002] Similar to visible light, X-rays are electromagnetic waves that can be divided into different energy spectra according to wavelength or frequency differences. X-rays with different energies have different attenuation characteristics, and these attenuation characteristics can reflect different physical properties of the materials under inspection. The traditional medical CT (Computed Tomography) detector adopts the X-ray energy integral detection method, integrates and receives X-ray photons of different energies, and reflects the average attenuation characteristics of X-rays. Therefore, the image reconstructed by medical CT is often difficult to obtain. Differentiate differences in imaging contrast of different soft tissues. In recent years, a new type of X-ray phot...

AI Technical Summary

Problems solved by technology

However, the consistency calibration method for the front-end by adjusting the gain of the amplifier requires multiple tests to determine the adjustment parameters, and when the product needs to be calibrated after leaving the factory, it still needs to be operated by the manufacturer, which increases the difficulty of maintenance and affects the subsequent analysis and processing.

In addition to the impact of the detector dead point on the consistency of the system response, the quantum noise generated during the X-ray photon detection process also has an impact on the consistency of the system response. The X-ray detected by the photon counting detector in a specific X-ray energy range The number of ray photons is limited, and the projection image contains more quantum noise, which has a great impact on the imaging effect. Therefore, a fast and efficient method is needed to eliminate the quantum noise in the projection image

Images