Method and apparatus for removing visible artefacts in video images

1. A method of removing structures of regularly occurring artifacts in decompressed video images, comprising the steps of:

providing a neural processor;

training the neural processor to specifically identify the structure of regularly occurring artifacts in a video signal;

providing a decompressed video signal;

filtering rows and columns of a video signal frame of the decompressed video signal frame-by-frame to extract data related to visible artifacts,

processing the extracted data frame-by-frame in the neural processor to produce data identifying the locations of the structures of the artifacts in the video signal; and

correcting the video signal to reduce the effect of identified structures of the artifacts.

2. The method according to claim 1 , in which the artifacts are identified over a number of frames of the video signal.

3. The method according to claim 1 , in which the artifacts are repeatedly identified.

4. The method according to claim 1 , in which the correcting step comprises filtering an image to reduce the effect of the artifacts.

5. The method according to claim 4 , in which the correcting step comprises applying a smoothing filter to the artifacts.

6. The method according to claim 1 , wherein the filtering is performed with a derivative filter.

7. The method according to claim 6 , wherein the neural processor detects spikes in the filtered data and determines which spikes relate to the visual artifacts.

8. The method according to claim 1 , wherein the neural processor utilizes a two-layer perception.

9. An apparatus for removing regularly occurring visible artifacts in a decompressed video signal, comprising:

means for receiving the video signal;

means for filtering the video signal frame-by-frame to extract data related to the artifacts;

a neural network processor trained and configured to specifically identify the artifacts and produce data relating to the location of the artifacts in the video signal; and

means for correcting the video signal by reducing the effects of the identified artifacts.

10. The apparatus according to claim 9 , wherein the neural network processor is configured to identify the artifacts over a number of frames.

11. The apparatus according to claim 10 , wherein the correcting means comprises means for filtering the image to reduce the effect of the artifacts.

12. The apparatus according to claim 11 , wherein the means for filtering comprises a smoothing filter.

13. The apparatus according to claim 9 , wherein the neural network processor is configured to repeatedly identify the artifacts.

14. A method for identifying regularly occurring visible artifacts in decompressed video images, comprising the steps of:

receiving a decompressed video signal;

filtering the video signal frame-by-frame to extract data relating to the artifacts;

training a neural network processor to specifically identify the artifacts;

identifying locations of the artifacts using the neural network processor; and

correcting the video signal to reduce the effect of the identified artifacts.

15. An apparatus for identifying regularly occurring visible artifacts in a decompressed video signal, comprising:

means for receiving a video signal;

means for filtering the video signal frame-by-frame to extract data related to the artifacts;

a neural network processor trained and configured to specifically identify the artifacts and produce data relating to the location of the artifacts in the video signal; and

means for correcting the video signal to reduce the effect of the identified artifacts.