Systems and methods for robust video temporal registration

1. A method of aligning frames between test and reference video sequences comprising the steps of:

distilling each frame of the test and reference video sequences into frame distillation measurements;

from the frame distillation measurements performing a linear alignment measurement using a linear Hough transform of a local Pearson's cross-correlation coefficient (LPCCC) image for spatial alignment of the frames between the test and reference video sequences to find a best fit line through the LPCCC image; and

for each pixel of the LPCCC image along the best fit line searching vertically for a higher correlation coefficient when the pixel has a value less than a threshold to find a better frame match for temporal alignment of the frames between the test and reference video sequences.

2. The method as recited in claim 1 wherein the frame distillation measurements comprise summaries of mean, standard deviation and RMS differences between corresponding frames of the test and reference video sequences.

3. The method as recited in claim 1 wherein the frame distillation measurements comprise summaries of mean, standard deviation and RMS differences between subimages for each pair of corresponding frames between the test and reference video sequences.

4. The method as recited in claim 1 where the frame distillation measurements comprise summary line and summary column one dimensional images of the LPCCC image.

5. The method as recited in claim 2 wherein the performing step comprises the steps of:

determining a best line fit mean value from the frame summary mean differences using a mean frame summary array pair as two one-by-framecount wide image inputs for the test and reference video sequences; and

comparing the best line fit mean value with the threshold.

6. The method as recited in claim 5 wherein the performing step further comprises the steps of:

if the best line fit means value is less than the threshold, determining a best line fit RMS value from the frame summary RMS differences using an RMS frame summary array pair as two one-by-framecount wide image inputs for the test and reference video sequences; and

comparing the best line fit RMS value with the threshold.

7. The method as recited in claim 6 wherein the performing step further comprises the steps of:

if the best line fit RMS value is less than the threshold, determining a best line fit standard deviation value from the frame summary standard deviation differences using a standard deviation summary array pair as two one-by-framecount wide image inputs for the test and reference video sequences; and

comparing the best line fit standard deviation value with the threshold.

8. The method as recited in claim 7 wherein the performing step further comprises the step of user selecting, if the best line fit standard deviation value is less than the threshold, the best line fit measurement from among the best line fit mean, RMS and standard deviation values having the highest correlation coefficient.

9. The method as recited in claim 3 wherein the distilling step comprises the steps of:

dividing each frame into image blocks; and

calculating corresponding mean, standard deviation and RMS differences per image block to provide the subimage summaries.

10. The method as recited in claim 9 further comprising the step of spatially aligning the subimages as in the performing step if the best correlation coefficient is less than the threshold to find a highest correlation coefficient per frame as a new frame match.

11. The method as recited in claim 4 further comprising the step of spatially aligning the summary line and summary column one dimensional images if the best correlation coefficient is less than the threshold to find a highest correlation coefficient per frame as a new frame match.