Image noise correction
AssigneeSTMicroelectronics S.r.l.
InventorsGrégory Roffet, Frédérique Crete
Life of this patent
Estimating the term…
Abstract
An image noise correction method is provided. For at least one target pixel having a determined pixel value, for each pixel in a window of pixels surrounding the target pixel, a weighting factor for the pixel is estimated based on the value of the target pixel and at least one pixel value in the window. An average of pixel values for the pixels in the window is calculated, with each pixel value being weighted by the weighting factor corresponding to the pixel. A new value is assigned to the target pixel based on the average of pixel values that is calculated. Also provided is an image noise correction device.
Claims20
A method for image noise correction, the method comprising, for at least one target pixel having a defined pixel value:
- for each pixel in a window of pixels surrounding the target pixel, estimating a weighting factor for that pixel, the weighting factor being based on the value of the target pixel and the value of at least one pixel within the window;
- calculating an average of pixel values for the pixels in the window, the calculating of the average of pixel values comprising weighting the value of each pixel within the window by the weighting factor that was estimated for that pixel;
- and
- assigning to the target pixel a new value that is based on the average of pixel values that was calculated,
- wherein the estimating of the weighting factor for each pixel in the window comprises determining a difference between the value of that pixel and the value of the target pixel.
A method for image noise correction, the method comprising, for at least one target pixel having a defined pixel value:
- for each pixel in a window of pixels surrounding the target pixel, estimating a weighting factor for that pixel based on the value of the target pixel and at least one pixel value within the window;
- calculating an average of pixel values for the pixels in the window, each pixel value being weighted by the weighting factor corresponding to the pixel;
- and
- assigning to the target pixel a new value that is based on the average of pixel values that is calculated,
- wherein the estimating of the weighting factor for each pixel in the window comprises, for each pixel in the window, estimating the corresponding weighting factor by inverting a distance between the value of that pixel and the value of the target pixel to which a constant has been added.
A method for image noise correction, the method comprising, for at least one target pixel having a defined pixel value:
- for each pixel in a window of pixels surrounding the target pixel, estimating a weighting factor for that pixel based on the value of the target pixel and at least one pixel value within the window;
- calculating an average of pixel values for the pixels in the window, each pixel value being weighted by the weighting factor corresponding to the pixel;
- and
- assigning to the target pixel a new value that is based on the average of pixel values that is calculated,
- wherein the estimating of the weighting factor for each pixel in the window comprises, for each pixel in the window, normalizing a distance between the value of that pixel and the value of the target pixel by taking into account a minimum distance and a maximum distance between the values of the pixels in the window and the value of the target pixel.
An image noise correction device comprising
- first memory for storing pixel values of pixels in a window around a target pixel;
- first processing means for estimating a weighting factor for each pixel in the window based on the pixel value of the target pixel and at least one pixel value in the window;
- second processing means for calculating an average of pixel values for the pixels in the window, each pixel value being weighted by the weighting factor corresponding to the pixel;
- and
- second memory for storing a new pixel value for the target pixel, the new value being based on the average of pixel values that is calculated by the second processing means.
A non-transitory computer-readable medium encoded with a computer program for image noise correction, the computer program comprising instructions for, for at least one target pixel having a defined pixel value:
- for each pixel in a window of pixels surrounding the target pixel, estimating a weighting factor for that pixel, the weighting factor being based on the value of the target pixel and the value of at least one pixel within the window;
- calculating an average of pixel values for the pixels in the window, the calculating of the average of pixel values comprising weighting the value of each pixel within the window by the weighting factor that was estimated for that pixel;
- and
- assigning to the target pixel a new value that is based on the average of pixel values that was calculated,
- wherein the estimating of the weighting factor for each pixel in the window comprises determining a difference between the value of that pixel and the value of the target pixel.
An image noise correction device comprising:
- first memory storing pixel values of pixels in a window around a target pixel;
- at least one processor programmed to:
- estimate a weighting factor for each pixel in the window based on the pixel value of the target pixel and at least one pixel value in the window;
- and
- calculate an average of pixel values for the pixels in the window, each pixel value being weighted by the weighting factor corresponding to the pixel;
- and
- second memory storing a new pixel value for the target pixel, the new value being based on the average of pixel values that is calculated by the at least one processor.
Dependencies are inferred from the claim text.