A filter network system and image processing system

Abstract

The invention discloses a filtering network system and an image processing system. The filtering network system includes: several Gaussian multipliers; an output circuit connected to several Gaussian multipliers respectively, wherein the Gaussian multiplier includes a first differential pair of tubes and The second differential pair of tubes and the third differential pair of tubes respectively connected to the first differential pair of tubes, the output terminals of the first transistors in the first differential pair of tubes are respectively connected to the input terminals of the two transistors in the second differential pair of tubes, the first differential pair of tubes The output ends of the second transistors in the pair are respectively connected to the input ends of the two transistors in the third differential pair, the input ends of the first transistor and the input ends of the second transistor are connected in parallel, and the output ends of the two transistors in the second differential pair are connected to the The output ends of the two transistors in the third differential pair are respectively connected to the output circuit. The invention realizes the Gaussian convolution filtering process by simulating the Gaussian multiplier, can improve the Gaussian convolution filtering speed and reduce the power consumption of the circuit.

Description

technical field [0001] The invention relates to the technical field of image processing, in particular to a filtering network system and an image processing system. Background technique [0002] In digital signal processing technology, convolution filtering is a very important step in signal preprocessing. In the field of image processing, the Gaussian convolution filter is a smoothing filter whose function is to obtain an image with a higher signal-to-noise ratio (SNR) and avoid image ringing. In a traditional digital image processing system, the image information collected by the image sensor is converted into a digital signal by an analog-to-digital converter ADC, and digital signal processing DSP technology is used for subsequent filtering processing. [0003] However, with the development of image sensing technology, the resolution of collected images has been greatly improved, making the continuous scale expansion of multi-scale Gaussian filtering lead to a sharp incr...

AI Technical Summary

Problems solved by technology

[0003] However, with the development of image sensing technology, the resolution of the collected images has been greatly improved, making the continuous scale expansion of multi-scale Gaussian filtering lead to a sharp increase in the amount of data and computation. It is becoming more and more difficult for traditional digital systems to meet the image Processing system requirements for real-time and power consumption

Especially embedded systems such as portable devices and wearable devices have higher design requirements in terms of high speed, miniaturization and ultra-low power consumption, and it is more difficult to meet the demand indicators

Images