Chinese Academy of Sciences

Beijing Smartlogic Technology Ltd.

Ruoshan Guo

Rui Han

Renjun Tang

Yang Luo

Fengli Yan

Xiaoli Tang

Lu Ye

Lei Wang

Donglin Wang

Chen Lin

The present invention discloses a contrast adaptive video denoising system, which comprises a frame memory for buffering the filtered frame; an inter-frame difference calculating module for inter-frame difference of the current input frame of the video and a previous filtered frame in the frame memory; a contrast calculating module for calculating the local contrast of the current input frame and inputting it into a low-contrast region detection module, a calculated low-contrast region confidence together with the inter-frame difference are input into a motion detection module to calculate the motion probability for each pixel. The motion adaptive temporal filtering module performs motion adaptive temporal filtering by using the current input frame of the video and the previous filtered frame in the frame memory as well as the motion probability of each pixel, and finally outputs the current filtered frame to store in the frame memory. Said system can solve the problems of motion tailing and blurring caused by conventional video denoising systems when processing low-contrast motion videos.

Contrast adaptive video denoising system

The present invention discloses an image anti-aliasing system, which comprises an edge detection unit, a directional filter, a low-angle directional filter, a directional filtering fusion unit and a result fusion unit; the edge detection unit outputs the edge direction and edge confidence according to an input image, the directional filter outputs a directional filtering result based on the input image and the edge direction, the low-angle directional filter outputs a low-angle directional filtering result according to the input image and the edge direction; the directional filtering fusion unit weights and combines the result output from the directional filter and the result output from the low-angle directional filter according to the edge direction, and outputs a result of weighted combination of two directional filtering results; the result fusion unit outputs an optimized image according to the edge confidence, the input image and the result of the directional filtering fusion unit. The present invention can well handle directional filtering in any direction and it has good effect to low-angle aliasing.

Image anti-aliasing system

The present invention provides apparatus and method for parallel polyphase image interpolation. The apparatus comprises: a local memory, a first memory access control unit, a second memory access control unit, a source image data buffering unit, a filter coefficient buffering unit, a multiply-accumulator, a third memory access control unit and a state machine; the first memory access control unit is configured for obtaining source image data from the local memory and buffering them into a source image data buffering unit; the second memory access control unit is configured for obtaining filter coefficients from the local memory and buffering them into a filter coefficient buffering unit; the source image data buffering unit is configured for inputting source image data to each multiply-accumulator, the filter coefficient buffering unit is configured for broadcasting filter coefficients corresponding to the source image data to the multiply-accumulators; and the multiply-accumulator is configured for performing multiply-accumulation operation on the source image data and the filter coefficients, and outputting the multiply-accumulation result every F clocks. The interpolation apparatus of the present invention is reconfigurable and can support interpolation filters with any order, so it is universal.

Apparatus and method for parallel polyphase image interpolation

The present invention discloses a video denoising system based on noise correlation, said system comprises a unit for estimating correlation of noises of adjacent pixels, which receives input of the inter-frame difference and motion probability, and estimates correlation of noises of adjacent pixels according to correlation of inter-frame differences between adjacent pixels in a still region, and outputs a noise correlation coefficient; a maximum filtering weight adjusting unit, which adaptively adjusts the maximum weight for temporal filtering according to the noise correlation coefficient and outputs the maximum weight for temporal filtering; the maximum weight for temporal filtering can control the range of fluctuation of the temporal filtering weight and the difference between the denoising effects for different pixels. The system of the present invention can solve the problem of 'speckle' noise occurred when the video noises have adjacent correlation in the conventional temporal denoising system for videos.

Video denoising system based on noise correlation

Growing community of inventors

Beijing, China

Ruoshan Guo