CMOS Image Sensor with Area-Efficient Correlated Double Sampling Scheme


Capacitive-type CMOS fingerprint sensors have advantages of compact-size and low-power consumption compared to optical sensors because they require no additional components such as light source and optical lens. Therefore, they are attractive to mobile applications. However, they suffer from image degradation for skin conditions known as dry and wet fingers. The other issue is the parasitic capacitances connected with the sensing electrode. These parasitic capacitances should be eliminated to reduce the required dynamic range of the readout circuitry and increase sensitivity. It is reported that the capacitively-coupled cells using pixel-parallel neuron-MOS logic circuits could give more accurate binarized image and the shield metal with its voltage controlled by an additional amplifier could reduce the parasitic capacitances.

In this work, a virtually-grounded metal shield effectively suppresses the parasitic capacitances without extra circuits and a configurable diffusion network is used to generate an adaptive local threshold level for pixel-level image enhancement [1]. A proposed 500dpi capacitive-type CMOS fingerprint sensor includes pixel-level image enhancement. The parasitic capacitances between finger skin and a sensing electrode are rejected to enhance the sensitivity. During the readout, capacitive diffusion networks generate a locally-smoothed average signal that can be used as a local threshold level. This local threshold signal is subtracted from the original sensing signal using analog circuits within the pixel. The output image is centered on the local threshold level resulting in better quality for binarization process.

Figure 1: CMOS fingerprint sensor with pixel level image enhancement


Related Publication
  1. Kwang-Hyun Lee, Euisik Yoon, "A 500dpi Capacitive-Type CMOS Fingerprint Sensor with Pixel-Level Adaptive Image Enhancement Scheme," Technical Digest of International Solid-State Circuits Conference, pp.282-283. Feb. 2002.