Unlock instant, AI-driven research and patent intelligence for your innovation.

Noise Suppression in Low Light Images

A noise suppression, low-light technology, applied in the field of image processing

Active Publication Date: 2016-11-16
MICROSOFT TECH LICENSING LLC
View PDF4 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

In many cases, common noise suppression algorithms can cause repetitive large speckle patterns to be generated in low light areas

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Noise Suppression in Low Light Images
  • Noise Suppression in Low Light Images
  • Noise Suppression in Low Light Images

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 100

[0021] Embodiment 100 illustrates a general purpose computer environment in which image processing can be performed. Image processing may include analyzing low-light images, or low-light portions of images, to jointly denoise and demosaic the images. The denoising part can use a technique where similar patches of the image are weighted, averaged, and a denoising value is determined for low light areas or for each patch in the image. The same set of similar patches and weights can be used for demosaicing to produce high definition and denoised images.

[0022] Image enhancement may be performed by device 102 . Device 102 is shown having hardware components 104 and software components 106 . The illustrated device 102 represents a conventional computing device, but other embodiments may have different configurations, architectures, or components.

[0023] In many embodiments, device 102 may be a personal computer or a server computer. In some embodiments, device 102 may also ...

Embodiment 200

[0040] Embodiment 200 may represent a method for denoising and demosaicing a raw image. The method of embodiment 200 may apply a weighted average of similar patches to determine a denoising value for each patch, and then use this same weighted average of similar patches when performing demosaicing. The method of embodiment 200 can skip similar patches that are not similar enough, and can skip the denoising process when sufficiently similar patches are not available.

[0041] In block 202, a raw digital image may be received. A raw digital image can be an image created with a single sensor and color filter array. A color filter array can filter individual pixels to a specific color. One such filtering mechanism may be a Bayer filter, where each block of four pixels is assigned two green pixels, one blue pixel, and one red pixel. From this image, the denoising process can reduce noise in low-light areas that may have a smaller signal-to-noise ratio.

[0042] After denoising,...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The present invention discloses noise suppression in low-light images. The low-light denoising mechanism may perform denoising prior to demosaicing, and may also use parameters determined during the denoising operation used to perform demosaicing. A denoising operation may attempt to find patches of the image that are similar to the first patch and use a weighted average based on the similarity to determine an appropriate value for denoising the original digital image. The same weighted average and similar slices can be used to demosaick the same image after the denoising operation.

Description

technical background [0001] The invention relates to image processing technology, in particular to noise suppression technology. Background technique [0002] Digital images are ubiquitous and can be captured by digital cameras, cell phones, video cameras, and many different devices. Many of the images may have been captured in low light conditions, or may have underexposed portions of the image. In many cases, common noise suppression algorithms can cause repetitive large speckle patterns to be generated in low light areas. Contents of the invention [0003] A scheme for low-luminosity noise reduction may perform denoising prior to demosaicing, and may use parameters determined during the denoising operation used to perform demosaicing. A denoising operation may attempt to find patches of the image that are similar to the first patch and use a weighted average based on the similarity to determine an appropriate value for denoising the original digital image. The same w...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Patents(China)
IPC IPC(8): G06T5/00
CPCG06T3/4015G06T5/70
Inventor P·柴特基N·乔希江胜明松下康之
Owner MICROSOFT TECH LICENSING LLC