Method and camera for detecting a region having a specific shape

1. A method of detecting a region having a specific shape by processing input image data, the method comprising:

calculating gradation differences between a central pixel and respective peripheral pixels in each of local regions of an image frame;

comparing an average gradation difference with each of the calculated gradation differences and obtaining local gradient pattern (LGP) values based on a comparison result in each of the local regions of the image frame; and

detecting the region having the specific shape from the image frame using the LGP values obtained from the respective local regions.

2. The method of claim 1 , wherein respective gradation data of a series of image frames having different sizes ranging from an image frame having a smallest size to an image frame having a greatest size are obtained,

the calculating gradation differences and the comparing an average gradation difference are performed on each of the image frames having different sizes to obtain LGP values of the respective image frames,

and the LGP values of the respective image frames are scanned using a sub-window in the detecting the region having the specific shape.

3. The method of claim 2 , wherein to determine whether or not any one region to which the sub-window is applied is the region having the specific shape in the detecting the region having the specific shape, when the sub-window is a non-shape region, learning result data of non-face weights of the LGP values are applied to specific positions of the sub-window.

4. The method of claim 3 , wherein the learning result data are a plurality of learning result data obtained when reference data of non-shape weights have different precisions, and learning result data having a lowest precision through learning result data having a highest precision are applied in a sequential order to any one region to which the sub-window is applied.

5. The method of claim 4 , wherein when any one of the learning result data is applied and it is determined that the any one region to which the sub-window is applied is the non-shape region, the non-face region is finally determined without applying the next learning result data,

and when it is determined that the any one region to which the sub-window is applied is the region having the specific shape, the next learning result data is applied.

6. The method of claim 5 , wherein when the learning result data having the highest precision is finally applied and it is determined that the any one region to which the sub-window is applied is the region having the specific shape, the region having the specific shape is finally determined.

7. A camera performing the method of claim 1 .

8. A method of detecting a region having a specific shape, the method comprising:

inputting image frame data;

obtaining gradation data of a series of image frames having different sizes;

obtaining local gradient pattern (LGP) values of each of the image frames having different sizes; and

detecting the region having the specific shape while scanning LGP values of each of the image frames using a sub-window.

9. The method of claim 8 , wherein, in the detecting the region having the specific shape, to determine whether any one region to which the sub-window is applied is the region having the specific shape, when the sub-window is a non-shape region, non-shape weights of the LGP values, are applied to specific positions of the sub-window.

10. The method of claim 8 , wherein the LGP values of each of the image frames having different sizes is obtained by calculating gradation differences between a central pixel and respective peripheral pixels in each of local regions of an image frame and comparing an average gradation difference with each of the calculated gradation differences and obtaining the LGP values based on a comparison result in each of the local regions.

11. A camera performing the method of claim 8 .