Method and system for measuring visual quality of a video sequence

1. A system comprising:

circuitry to:

identify at least one spatial temporal human visual system model visibility setting;

apply at least one spatial temporal human visual system model visibility setting to a reference video sequence to generate a modified reference video sequence;

apply the at least one spatial temporal human visual system model visibility setting to a distorted video sequence to generate a modified a distorted video sequence;

a quality determination unit to generate a quality assessment score by determining a difference metric between a decorrelated transform domain representation of the reference video sequence and a decorrelated transform domain representation of the distorted video sequence following application of a set human visibility quantizers to the decorrelated transform domain representation of the distorted video sequence and the decorrelated transform domain representation of the reference video sequence.

2. The system as recited in claim 1 , wherein the at least one spatial temporal human visual system model visibility setting includes human visibility quantizers applicable to a decorrelated transform domain.

3. The system as recited in claim 1 , wherein the quality assessment score is based at least in part on an error metric or a difference metric.

4. The system as recited in claim 1 , wherein the quality assessment score is based at least in part on a mean squared error calculation.

5. The system as recited in claim 1 , wherein the quality assessment score is a distortion metric determined based at least in part on the difference metric.

6. The system as recited in claim 1 , further comprising a compressor configured to receive the quality assessment score and to adjust at least one parameter associated with operations performed with respect to generating the distorted video sequence.

7. The system as recited in claim 1 , wherein the distorted video sequence is a pre-processed version of the reference video sequence.

8. The system as recited in claim 1 , wherein the distorted video sequence is a compressed and decompressed version of the reference video sequence.

9. A method comprising:

applying at least one spatial temporal human visual system model visibility setting to a reference video sequence to generate a modified reference video sequence, the at least one spatial temporal human visual system model visibility setting includes human visibility quantizers applicable to a decorrelated transform domain, wherein applying the at least one spatial temporal human visual system model visibility setting to the reference video sequence includes applying a set of human visibility quantizers to a decorrelated transform domain representation of the reference video sequence;

applying the at least one spatial temporal human visual system model visibility setting to a distorted video sequence to generate a modified distorted video sequence, wherein applying the at least one spatial temporal human visual system model visibility setting to the distorted video sequence includes applying the set of human visibility quantizers to a decorrelated transform domain representation of the distorted video sequence; and

generating a quality assessment score based at least in part on the modified reference video sequence and the modified distorted video sequence, wherein generating the quality assessment score includes:

determining a difference metric between the modified reference video sequence and the modified reference video sequence within the decorrelated transform domain; and

determining a distortion metric based at least in part on the difference metric.

10. The method as recited in claim 9 , further comprising adjusting at least one parameter associated with operations performed with respect to generating the distorted video sequence.

11. The method as recited in claim 10 , wherein the operations are associated with pre-processing the reference video sequence to generate the distorted video sequence.

12. The system as recited in claim 10 , wherein the operations are associated with compressing the reference video sequence to generate the distorted video sequence.

13. One or more non-transitory computer-readable media having computer-executable instructions which, when executed by one or more processors, cause the one or more processors to perform operations comprising:

applying a set of human visibility quantizers to a decorrelated transform domain representation of a reference video sequence;

applying the set of human visibility quantizers to a decorrelated transform domain representation of a distorted video sequence;

determining a difference metric between the decorrelated transform domain representation of the reference video sequence and the decorrelated transform domain representation of the distorted video sequence, following application of the set human visibility quantizers to the decorrelated transform domain representation of the distorted video sequence and the decorrelated transform domain representation of the reference video sequence; and

determining a quality assessment score based at least in part on the difference metric.

14. The one or more non-transitory computer-readable media as recited in claim 13 , wherein the operations further comprise generating the distorted video sequence by performing at least one operation on the reference video sequence.

15. The one or more non-transitory computer-readable media as recited in claim 13 , wherein the quality assessment score is based at least in part on a mean squared error calculation.

16. The one or more non-transitory computer-readable media as recited in claim 13 , wherein the quality assessment score is based at least in part on a distortion metric.

17. The one or more non-transitory computer-readable media as recited in claim 13 , wherein the set of human visibility quantizers are representative of at least one viewing condition associated with displaying the distorted video sequence.

18. The one or more non-transitory computer-readable media as recited in claim 13 , wherein the quality assessment score is based at least in part on a mean squared error of various block sizes of the set of human visibility quantizers associated with each block of pixels in the decorrelated transform domain representation of the reference video sequence and the decorrelated transform domain representation of the distorted video sequence.

19. The one or more non-transitory computer-readable media as recited in claim 13 , further comprising providing the quality assessment score to an encoder to assist in tuning a pre-processing of the distorted video sequence.