System for reducing noise in video processing
A system includes a data storage configured to store a model human visual system, an input module configured to receive an original picture in a video sequence and to receive a reference picture, and a processor. The processor is configured to create a pixel map of the original picture using the model human visual system. A first layer is determined from the pixel map. A weighting map is determined from a motion compensated difference between the original picture and the reference picture. A processed picture is then determined from the original picture using the weighting map and the first layer.
1. A system for reducing noise in video processing, the system comprising:
a data storage configured to store a model human visual system;
an input module configured to receive an original picture in a video sequence and to receive a reference picture in the video sequence;
a processor configured to:
create a pixel map of the original picture using the model human visual system;
determine a first layer from the pixel map using the model human visual system;
determine a motion compensated difference between the original picture and the reference picture,
determine a weighting map from the motion compensated difference between the original picture and the reference picture, wherein the weighting map includes a value for each pixel based on a model of human temporal perceptibility; and
determine a processed picture from the original picture by recombining the first layer with the weighting map,
wherein the first layer goes beyond a noise layer that includes additive noise and is made up of a spatial detail layer including a value for each pixel based on a model of human spatial perceptibility, and
wherein a three dimensional noise reducer (3DNR) process precedes an adaptive detail reducer (ADR) process, and wherein the 3DNR performs three dimensional noise reduction on the original picture in two spatial dimensions and a temporal dimension, and the ADR process performs adaptive detail reduction on the original picture on selected difficult-to-track details.
2. The system of claim 1 , wherein the difficult-to-track details in the original picture may be reduced using the weighting map and the ADR to form the processed picture 130 .
3. The system of claim 2 , wherein the difficult-to-track details may be determined using a predetermined threshold based on the weighting map.
4. The system of claim 3 , wherein the model human visual system takes into account statistics of natural images and a response function of cells in a retina.
5. The system of claim 4 , wherein an asymmetric (infinite impulse response) IIR is used determine the weighting map.
6. A system for reducing noise in video processing, the system comprising:
a data storage configured to store a model human visual system;
an input module configured to receive an original picture in a video sequence and to receive a reference picture in the video sequence;
a processor configured to:
create a pixel map of the original picture using the model human visual system;
determine a first layer from the pixel map using the model human visual system;
determine a motion compensated difference between the original picture and the reference picture,
determine a weighting map from the motion compensated difference between the original picture and the reference picture, wherein the weighting map includes a value for each pixel based on a model of human temporal perceptibility; and
determine a processed picture from the original picture by recombining the first layer with the weighting map,
wherein the first layer goes beyond a noise layer that includes additive noise and is made up of a spatial detail layer including a value for each pixel based on a model of human spatial perceptibility, and
wherein an asymmetric (infinite impulse response) IIR is used determine the weighting map.