A square root analog to digital converter to optimally convert photonic signals for computed tomography

Abstract

The arrival of photons at a given location is a Poisson process with an associated shot noise which rises with the square root of the number of photons received. An analog-to-digital converter (ADC) with a square root transfer function can quantize photonic signals with LSB size kept constant with respect to the photon shot noise. In imaging applications, this can greatly reduce the number of bits needed to characterize a signal compared to a linear ADC without detrimental effects to image quality. Such a device, based on the Analogic MuSIC chip, was designed and tested for the needs of a medical computed tomography (CT) device. The experimental setup increases the MuSIC sampling frequency from 3kHz to 7kHz, while reducing the amount of data necessary for reconstruction. A constant quantization noise to photon shot noise ratio acceptable for CT is maintained by sizing each LSB to be one half the rms noise level. Results show an INL of 2.5 LSB, which is reduced to 0.27 LSB after a correction scheme.