Scene change detection
View Patent ↗Scene change detection in encoding digital pictures is disclosed. A statistical quantity μ M is calculated for a given section in a current picture. A window of one or more sections is defined around a co-located section in a previous picture. A statistical sum E is calculated over the sections in the window. A difference between the statistical sum E and the statistical quantity μ M is calculated. The difference between E and μ M is used to determine whether the given section is a scene-change section. Whether the current picture is a scene-change picture may be determined from the number of scene change sections. Information indicating whether or not the current picture is a scene-change picture may be stored or transferred.
1. A method for detecting a scene change in encoding one or more digital pictures, comprising:
a) calculating a statistical quantity for a given section in a current picture;
b) defining a window of one or more sections around a co-located section in a previous picture;
c) calculating a statistical sum over the one or more sections in the window;
d) calculating a difference between the statistical sum and the statistical quantity for the given section;
e) using the difference between the statistical sum and the statistical quantity for the given section to determine whether the given section is a scene-change section;
f) determining a number of scene-change sections for the current picture;
g) determining whether the current picture is a scene-change picture from the number of scene change sections; and
h) storing or transferring information indicating whether or not the current picture is a scene-change picture,
wherein c) includes determining a weight for each section in the window, wherein the statistical sum is a weighted sum given by a sum of products, wherein each product in the statistical sum is a product of a weight and a corresponding statistical value for a given section of the one or more sections, wherein the statistical sum is taken over all sections in the window.
2. The method of claim 1 , further comprising:
i) encoding the current picture with an encoder according to the information indicating whether or not the current picture is a scene-change picture.
3. The method of claim 2 , wherein i) includes encoding the current picture as an I-frame if the information indicates that the current picture is a scene-change picture or encoding the current picture as a B-frame or P-frame if the information indicates that the current picture is not a scene-change picture.
4. The method of claim 1 , further comprising, repeating a) through e) for a plurality of sections in the current picture.
5. The method of claim 1 , wherein each section is a macroblock.
6. The method of claim 1 , wherein a) includes calculating a mean pixel value or a variance of a pixel value for each section in the current picture.
7. The method of claim 1 , wherein the window includes one or more sections neighboring the co-located section.
8. The method of claim 1 , wherein determining the weight for each section in the window includes determining a movement direction and assigning each weight according to a proximity of the corresponding section in the window to the movement direction.
9. The method of claim 1 , wherein e) includes comparing the difference between the statistical sum and the statistical quantity to a threshold value.
10. The method of claim 1 , wherein g) includes comparing the number of scene change sections to a threshold.
11. The method of claim 1 , wherein g) includes determining whether the number of scene-change sections is a majority of sections in the current picture.
12. A system for detecting a scene change in encoding one or more digital pictures, comprising:
a processor;
a memory coupled to the processor;
instructions embodied in the memory and executable by the processor, wherein the instructions are configured to implement a method for detecting a scene change in encoding the one or more digital pictures, the method comprising:
a) calculating a statistical quantity for a given section in a current picture;
b) defining a window of one or more sections around a co-located section in a previous picture;
c) calculating a statistical sum over the one or more sections in the window;
d) calculating a difference between the statistical sum and the statistical quantity for the given section;
e) using the difference between the statistical sum and the statistical quantity for the given section to determine whether the given section is a scene-change section;
f) determining a number of scene-change sections for the current picture;
g) determining whether the current picture is a scene-change picture from the number of scene change sections; and
h) storing or transferring information indicating whether or not the current picture is a scene-change picture,
wherein c) includes determining a weight for each section in the window, wherein the statistical sum is a weighted sum given by a sum of products, wherein each product in the statistical sum is a product of a weight and a corresponding statistical value for a given section of the one or more sections, wherein the statistical sum is taken over all sections in the window.
13. The system of claim 12 , further comprising an encoder configured to encode the current picture according to the information indicating whether or not the current picture is a scene-change picture.
14. The system of claim 12 , wherein the instructions are further configured to repeat a) through e) for a plurality of sections in the current picture.
15. The system of claim 12 , wherein determining the weight for each section in the window includes determining a movement direction and assigning each weight according to a proximity of the corresponding section in the window to the movement direction.
16. The system of claim 13 , wherein the encoder is configured to encode the current picture as an I-frame if the information indicates that the current picture is a scene-change picture or encoding the current picture as a B-frame or P-frame if the information indicates that the current picture is not a scene-change picture.
17. A non-transitory computer-readable storage medium having computer program instructions embodied therein, wherein the computer program instructions are configured, when executed, to implement a method for detecting a scene change in encoding one or more digital pictures, the method comprising:
a) calculating a statistical quantity for a given section in a current picture;
b) defining a window of one or more sections around a co-located section in a previous picture;
c) calculating a statistical sum over the one or more sections in the window;
d) calculating a difference between the statistical sum and the statistical quantity for the given section;
e) using the difference between the statistical sum and the statistical quantity for the given section to determine whether the given section is a scene-change section;
f) determining a number of scene-change sections for the current picture;
g) determining whether the current picture is a scene-change picture from the number of scene change sections; and
h) storing or transferring information indicating whether or not the current picture is a scene-change picture,
wherein c) includes determining a weight for each section in the window, wherein the statistical sum is a weighted sum given by a sum of products, wherein each product in the statistical sum is a product of a weight and a corresponding statistical value for a given section of the one or more sections, wherein the statistical sum is taken over all sections in the window.
18. The non-transitory computer-readable storage medium of claim 17 , wherein the computer program instructions are further configured to encode the current picture according to the information indicating whether or not the current picture is a scene-change picture.