Correlated multiple sampling technique-a discrete Fourier Transform analysis aimed for CMOS image sensors

Modern CMOS image sensors generally face difficulties in limiting the noise effect of their readout circuits. The existing noise sources range from spatial to temporal noise, such as the thermal, the flicker, and the random telegraph signal noise sources. The noise sources and the device's image quality issues become more evident when targeting design goals, such as high resolution and high frame-rate. This paper focuses on the theoretical derivation of the Correlated Multiple Sampling technique by means the Discrete Fourier Transform in which it explicitly includes the noise contamination process during the image sensor readout operation. The purpose of the proposed approach is to deepen the underlying multiple sampling process role in low noise systems and how beneficial this technique is in suppressing the noise contamination, as such proposing a more intuitive way to understand the method.