Noise Suppression in Low Light Images
A noise suppression, low-light technology, applied in the field of image processing
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
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,...
PUM
Login to View More Abstract
Description
Claims
Application Information
Login to View More 