abstract |
The present application proposes an ultra-high-speed X-ray image recognition method and device based on a SiC neural network chip. The method includes using a SiC detector to construct a SiC neural network chip similar to the visual nervous system of the human eye, so that the SiC neural network chip has The ultra-high-speed X-ray image perception capability, after being trained on the X-ray image data set, constitutes the device; the device specifically includes: a SiC neural network chip, and the SiC neural network chip is a specially arranged SiC array detector, It is used for receiving X-rays and converting the X-rays into electrical signals; the pixel weight modulation unit provides a reverse bias voltage for the SiC array detector, so that each pixel of the SiC array detector has an effect on X-rays. Different responsiveness; a signal readout system that receives and processes the electrical signal and outputs a signal matrix that represents the result of image recognition. The method and the device are suitable for ultra-high-speed identification of X-ray images in a strong radiation field, and can break through the speed limit of the existing X-ray image identification methods and devices. |