Combining image sensing and Gabor-type filtering in analog VLSI

Abstract

We describe the design, implementation and test results from an analog CMOS VLSI cellular neural network chip which integrates photosensing and an image filtering operation similar to Gabor-filtering. The design reported senses a one-dimensional 27 pixel input image focused onto the chip and generates three outputs at every pixel: the input image intensity, the real part of the filter output, and the imaginary part of the filter output. Test results indicate that the chip functions as expected.