Method for generating an image and an image generator

1. A reference image generator, comprising:

a memory for storing a second image that is acquired during a second exposure period and a reference image; and

an image processor, adapted to:

(i) define a first image as a reference image, wherein the first image is acquired during a first exposure period that precedes the second exposure period;

(ii) define multiple reference regions of interest within the reference image, wherein an aggregate number of pixels within the multiple reference regions of interest is substantially smaller than a number of pixels of the first image; and

(iii) update the reference image in response to content of the reference image, content of the second image and a first type of spatial relationship defined between the reference image and the second image; wherein a shift is defined in response to multiple spatial relationships between the multiple reference regions of interest and between multiple regions of interest of the second image.

2. The image generator according to claim 1 , wherein the memory is adapted to store a new image that is acquired during a new exposure period; and the image processor is adapted to update the reference image in response to content of the new image, content of the reference image and an inter-image spatial relationship defined between the new image and the reference image.

3. The image generator according to claim 1 , wherein the image processor is adapted to define non-overlapping regions of interest.

4. The image generator according to claim 1 , wherein the image processor is adapted to:

find multiple points of interest, wherein a point of interest is characterized by a high spatial derivative value; and

select multiple regions of interest, wherein a region of interest includes a point out of the group of points and wherein the selecting is responsive to at least one auto-correlation value of the region of interest.

5. The image generator according to claim 4 , wherein the image processor is adapted to calculate multiple correlation values between a certain region of interest and multiple spatially shifted versions of the certain region of interest.

6. The image generator according to claim 5 , wherein the image processor is adapted to calculate multiple spatial shifted versions wherein the spatial shift is responsive to a Bayer pattern of the first image.

7. The image generator according to claim 4 , wherein the image processor is adapted to ignore points of interest that are relatively proximate to each other.

8. The image generator according to claim 1 , wherein the image processor is adapted to define at least one region of interest within a background portion of the reference image and define at least one region of interest within a foreground portion of the reference image.

9. The image generator according to claim 1 , wherein the image processor is adapted to:

generate a gray level representation of the multiple reference regions of interest,

generate a gray level representation of the multiple regions of interest of the second image; and

calculate the multiple spatial relationships between the multiple reference regions of interest and between multiple regions of interest of the second image in response to correlations between gray level representations of reference regions of interest and corresponding regions of interest of the second image.

10. The image generator according to claim 9 , wherein the image processor is adapted to calculate the shift in response to similarity between the multiple spatial relationships.

11. The image generator according to claim 9 , wherein the similarity is responsive to locations of maximal valued elements within region of interest correlation matrices.

12. The image generator according to claim 1 , wherein the image processor is adapted to further calculate a fixed noise distribution and to update the reference image in response to the fixed noise distribution.

13. The image generator according to claim 1 , wherein the image processor is adapted to up-scale the first image and the second image before updating the reference image.

14. A computer program product of a non-transitory computer readable medium usable with a programmable computer, the computer program product having computer-readable code embodied therein for generating a reference image, the computer-readable code comprising instructions for:

receiving a first image that is acquired during a first exposure period and defining the first image as a reference image;

defining multiple reference regions of interest within the reference image, wherein an aggregate number of pixels within the multiple reference regions of interest is substantially smaller than a number of pixels of the first image;

receiving a second image that is acquired during a second exposure period; and

updating the reference image in response to a content of the reference image, content of the second image and a first type of spatial relationship defined between the reference image and the second image,

wherein a shift is defined in response to multiple spatial relationships between the multiple reference regions of interest and between multiple regions of interest of the second image.

15. The computer program product of claim 14 , wherein the instructions for generating an image comprise instructions for receiving a new image that is acquired during a new exposure period; and updating the reference image in response to content of the new image, content of the reference image and a further shift between the new image and the reference image.