Detecting upscaled source video

A method comprising:

• estimating a native size of frames from a video sequence by, for each of a plurality of frames from the video sequence:
• converting the respective frame to a domain of frequency coefficients, each coefficient representing content of the input frame as a respective pair of frequencies each extending in one of two dimensions;
• for each converted frame:
• for each set of frequency coefficients having a common frequency in a first dimension considered in order by frequency in a second dimension, identifying zero crossings among the set of frequency coefficients,
• counting the zero crossings from among the sets at each frequency in the second dimension, and
• estimating whether the input frame was upsampled from the count of zero crossings;
• and
• when the number of input frames that are estimated as being upsampled exceed a predetermined value, rejecting the video sequence.

A non-transitory computer readable medium storing program instructions that, when executed by a processing device, causes the device to:

• estimate a native size of frames from a video sequence by, for each of a plurality of frames from the video sequence:
• converting each frame to a domain of frequency coefficients, each coefficient representing content of the respective frame as a respective pair of frequencies each extending in one of two dimensions, for each converted frame:
• for each set of frequency coefficients having a common frequency in a first dimension considered in order by frequency in a second dimension, identifying zero crossings among the set of frequency coefficients, counting the zero crossings from among the sets at each frequency in the second dimension, and
• estimating whether the input frame was upsampled from the count of zero crossings;
• and when the number of selected frames that are estimated as being upsampled exceeds a predetermined value, reject the video sequence.