Video quality abnormality detection method and apparatus

• US 6,377,299 B1
• Filed: 03/24/1999
• Issued: 04/23/2002
• Est. Priority Date: 04/01/1998
• Status: Expired due to Term

• Current Assignee: KDDI Corporation
• Original Assignee: KDD Corporation (KDDI Corporation)
• Inventors: Hamada, Takahiro
• Primary Examiner(s): Garber, Wendy R.
• Assistant Examiner(s): DESIR, JEAN WICEL
• Application Number:
  US09/274,675
• Time in Patent Office: 1,126 Days
• Field of Search: H04/N 1.700, H04/N.521, 348/192, 348/189, 348/180, 348/184, 348/185, 348/715, 348/554, 348/555, 348/607, 348/608, 348/627, 348/909, 382/239, 382/278, 382/280, 382/286, 382/191, 382/274, 382/275, 358/533, 358/504, 386/21, 386/76, 386/113, 386/114
• US Class Current: 348/192
• CPC Class Codes: H04N 17/00 Diagnosis, testing or measuring for television systems or their details

Abstract

A waveform analysis unit divides digital data of a video signal stored in a Y memory and a C memory into small blocks, and conducts average value separation and normalization on each of the small blocks. A 1/f fluctuation measuring unit first derives a self-correlation function for each of the normalized small blocks, folds back the self-correlation function to the negative side, and conducts correction processing for making an average self-correlation function over a predetermined section equal to 0. Subsequently, the 1/f fluctuation measuring unit conducts Fourier transform on the corrected self-correlation function, derives a power spectrum, and conducts straight line approximation of the power spectrum. The 1/f fluctuation measuring unit then derives a fluctuation value of the power spectrum from the approximate straight line. An abnormality determination unit makes a determination on video quality abnormality on the basis of the variance value of the small block, coefficients of the approximate straight line derived by the straight line approximation, and the fluctuation value.

Claims

1. A video quality abnormality detection method comprising the steps of:

dividing a digitized video signal into a plurality of blocks;

normalizing an original signal with each of divided blocks, as a unit, so as to make a power sum of an AC component constant;

deriving a power spectrum on the basis of data resulting from the normalizing;

deriving a fluctuation value from an approximate straight line of the power spectrum and the power spectrum; and

detecting the video quality abnormality from at least a variance of the signal of the block, coefficients of the approximate straight line, and the fluctuation value of the power spectrum.

2. A video quality abnormality detection apparatus comprising:

storage means for storing a digitized luminance signal and a color-difference signal;

variance value computation means for dividing the luminance signal and the color-difference signal stored in the storage means into blocks, and deriving variance values of each block;

normalization means for deriving normalized signals each having a constant variance value, on the basis of the variance values derived by the variance value computation means;

self-correlation function correction processing means for deriving average self-correlation functions within the block from the normalized signals outputted from the normalization means, folding back each of the average self-correlation functions to a negative side, and making the average self-correlation functions over a predetermined section equal to 0 or nearly 0;

power spectrum straight line approximation means for conducting Fourier transform on the self-correlation functions derived by the self-correlation function correction processing means, deriving power spectra, and approximating the power spectra by straight lines;

fluctuation value computation means for deriving fluctuation values of the power spectra from the approximate straight lines; and

a means for making a determination on video abnormality on the basis of the variance values derived by the variance value computation means, coefficients of the approximate straight lines derived by the power spectrum straight line approximation means, and the fluctuation values derived by the fluctuation value computation means.

3. A video quality abnormality detection apparatus according to claim 2, wherein the normalization means divide the original signals obtained from the storage means by the variance values.

4. A video quality abnormality detection apparatus according to claim 2, wherein the power spectrum straight line approximation means approximate each of the power spectra by “

a log f+b”

, where a and b are coefficients, and f is a frequency.

5. A video quality abnormality detection apparatus according to claim 4, wherein the fluctuation value computation means add together absolute values of difference between each of the power spectra and the straight line “

a log f+b”

over a predetermined section of the frequency f.