Initial advancement of image sensor IC technology from the 1970s to 1990s focused on improving signal to noise ratio or detector limited performance. This was achieved by extending integration time of each pixel to the complete available exposure time and reducing readout noise below the detector dark noise. First phase of pixel miniaturization helped convert scanning imagers into time-delay integration (TDI) scanners and eventually into staring arrays, thus extending their integration time. The read noise was reduced through input circuit innovations such as source follower circuit, capacitive trans-impedance amplifier (CTIA), gate-modulation (GM), direct-injection (DI), buffered direct-injection (BDI) circuit and other more advanced circuits.
The second wave of image sensor IC innovation in the last two decades focused on developing large formats with smaller pixels. To regain low noise performance on large focal plane arrays (image sensor sensors) without compromising on frame rates the analog ICs needed to be converted into digital ICs. Digital ICs started from having digitizer (ADC) per chip in late 1990s, to ADC per column in 2000s and eventually per pixel ADC were developed. Other innovative IC technology was also demonstrated in the past two decades. These included multi-color pixels, pixels with intensity and ranging capabilities, ultra-low noise astronomy pixels supporting very long integration, and rad-hard by design ICs for space. These image sensor IC innovations of the past have been enabled by shrinking CMOS design rules and other CMOS process improvements.
