Techniques for evaluating compressed motion video quality

Patent metadata

Field	Value
Publication number	US-9992500-B2
Application number	US-201414218352-A
Country	US
Kind code	B2
Filing date	Mar 18, 2014
Priority date	Mar 18, 2014
Publication date	Jun 5, 2018
Grant date	Jun 5, 2018

Abstract

Various embodiments are generally directed to techniques for evaluating video quality of compressed versions of a motion video to select compressed frames of that motion video without access to an uncompressed version. A device to transmit motion video includes a device scoring component to select a set of coefficients from a vector correlating temporal complexity values to sets of coefficients based on a temporal complexity of a first compressed frame of a first compressed video data, the vector derived from opinion scores associated with at least one viewing characteristic of a viewing device; and a selection component to select either the first compressed frame or a second compressed frame of a second compressed data to transmit to the viewing device based on a metric of video quality derived from the selected set of coefficients, the first and second compressed data representing a motion video. Other embodiments are described and claimed.

First claim

The invention claimed is:

1. A device to transmit motion video comprising:
a device scoring component to select a set of coefficients from a plurality of vectors that each correlate temporal complexity values to one or more sets of coefficients based on a temporal complexity of a first compressed frame of a first compressed video data, the plurality of vectors each derived from opinion scores associated with a different set of viewing characteristics of at least one viewing device;
a selection component to select one of the first compressed frame and a second compressed frame of a second compressed video data to transmit to a viewing device with at least one viewing characteristic associated with the selected set of coefficients based on a metric of video quality derived from the selected set of coefficients, the first and second compressed video data representing a motion video; and
a buffering component to transmit an indication to the viewing device to alter a degree of buffering provided by the viewing device to enable viewing of the motion video on a display of the viewing device based on the selection of one of the first and second compressed frames.

2. The device of claim 1 , comprising a decoder to decompress the first compressed frame to derive at least one parameter of the first compressed frame, the at least one parameter comprising at least one of a quantization parameter (QP) employed by a compressor to generate the first compressed frame, a resolution of the first compressed frame, a type of the first compressed frame, a data size of the first compressed frame or a bitrate of the first compressed frame.

3. The device of claim 2 , comprising a complexity component to derive the temporal complexity of the first compressed frame from the at least one parameter.

4. The device of claim 3 , comprising a mean opinion score (MOS) estimator to use at least the selected set of coefficients and a bitrate of the first compressed frame to derive an estimated MOS of the first compressed frame, the metric of video quality comprising the estimated MOS.

5. The device of claim 3 , comprising a bitrate estimator to use at least the selected set of coefficients and a desired MOS value to derive an estimated bitrate required to achieve the desired MOS value, the metric of video quality comprising the estimated bitrate.

6. The device of claim 1 , comprising an interface to transmit the motion video in a compressed form comprising the selected one of the first and second compressed frames to the viewing device.

7. A device to transmit motion video comprising:
a non-linear fitting component to derive from raw data a plurality of mathematical models that each correlate a range of temporal complexities to a range of opinion scores associated with a different set of viewing characteristics of at least one viewing device, the raw data comprising opinion scores collected from using at least one viewing device associated with each set of viewing characteristics to view multiple motion videos of different temporal complexities of a training set of motion videos;
a linear fitting component to derive at least one vector for each of the plurality of mathematical models that correlates temporal complexity values to sets of coefficients associated with each of the different sets of viewing characteristics; and
a buffering component to transmit an indication to the viewing device to alter a degree of buffering provided by the viewing device to enable viewing of the motion video on a display of the viewing device based on at least the at least one vector.

8. The device of claim 7 , comprising a collection component to monitor controls for an indication of operation of the controls to provide at least one of the opinion scores.

9. The device of claim 8 , the collection component to provide the multiple motion videos to a viewing device associated with a selected set of viewing characteristics.

10. The device of claim 7 , the multiple motion videos of the training set selected to provide the training set with a selected range of temporal complexities with a selected distribution of temporal complexities.

11. The device of claim 7 , comprising an interface to transmit the at least one vector to another device to enable the other device to use the at least one vector with a temporal complexity of another motion video to control selection of compressed frames representing the other motion video to another viewing device having the at least one viewing characteristic.

12. A computer-implemented method for transmitting motion video comprising:
selecting a set of coefficients from a plurality of vectors that each correlate temporal complexity values to one or more sets of coefficients based on a temporal complexity of a first compressed frame of a first compressed video data, the plurality of vectors each derived from opinion scores associated with a different set of viewing characteristics of at least one viewing device;
selecting one of the first compressed frame and a second compressed frame of a second compressed video data to transmit to a viewing device with at least one viewing characteristic associated with the selected set of coefficients based on a metric of video quality derived from the selected set of coefficients, the first and second compressed video data representing a motion video; and
transmitting an indication to the viewing device to alter a degree of buffering provided by the viewing device to enable viewing of the motion video on a display of the viewing device based on the selection of one of the first and second compressed frames.

13. The computer-implemented method of claim 12 , the method comprising decompressing the first compressed frame to derive at least one parameter of the first compressed frame, the at least one parameter comprising at least one of a quantization parameter (QP) employed in generating the first compressed frame, a resolution of the first compressed frame, a type of the first compressed frame, a data size of the first compressed frame or a bitrate of the first compressed frame.

14. The computer-implemented method of claim 13 , the method comprising deriving a temporal complexity of the first compressed frame from the at least one parameter.

15. The computer-implemented method of claim 14 , the method comprising using at least the selected set of coefficients and a bitrate of the first compressed frame to derive an estimated MOS of the first compressed frame, the metric of video quality comprising the estimated MOS.

16. The computer-implemented method of claim 12 , the opinion scores associated with viewing multiple motion videos of a training set of motion videos using viewing devices having selected sets of viewing characteristics.

17. The computer-implemented method of claim 12 , the at least one viewing characteristic comprising at least one of a display size, a display resolution or a viewing distance.

18. The computer-implemented method of claim 12 , the method comprising transmitting the motion video in a compressed form comprising the selected one of the first and second compressed frames to the viewing device.

19. At least one non-transitory machine-readable storage medium comprising instructions that when executed by a computing device, cause the computing device to:
select a set of coefficients from a plurality of vectors that each correlate temporal complexity values to one or more sets of coefficients based on a temporal complexity of a first compressed frame of a first compressed video data, the plurality of vectors

Assignees

• Intel Corp

Inventors

• Choi Lark Kwon
• Liao Yiting
• O'Mahony Barry

Classifications

• H04N21/44004

  involving video buffer management, e.g. video decoder buffer or video display buffer · CPC title

• H04N19/154

  Measured or subjectively estimated visual quality after decoding, e.g. measurement of distortion (use of rate-distortion criteria H04N19/147) · CPC title

• H04N21/6543

  for forcing some client operations, e.g. recording {(remote booting in general G06F9/4416)} · CPC title

• H04N19/172 Primary

  the region being a picture, frame or field · CPC title

• H04N19/12

  Selection from among a plurality of transforms or standards, e.g. selection between discrete cosine transform [DCT] and sub-band transform or selection between H.263 and H.264 · CPC title