Method of detecting a specific object in an image signal

1. A method of detecting a specific object in an image signal comprising:

detecting an initial object region from an input image using a general detection algorithm if an object region has not been detected in a previous input image; and

detecting an object region from an input image by a template matching based upon an effective range if an object region was detected in a previous input image.

2. The method of claim 1 , wherein detecting an initial object region from an input image using a general detection algorithm comprises detecting the initial object region by either a template matching method using a current input image, or a method using an image difference and color information.

3. The method of claim 1 , wherein detecting an object region from an input image by a template matching based upon an effective range comprises defining the effective range as an estimated range where an object region may appear in a next input image.

4. The method of claim 1 , wherein detecting an object region from an input image by a template matching based upon an effective range comprises performing the template matching by at least one of a whole matching method based on template matchings in predetermined positions, a motion estimation matching method based on a motion of an object, and a speed estimation matching method based on a relative speed difference of object motion appearing in the input images during a defined time interval (Δt).

5. The method of claim 4 , wherein the template matching in the whole matching method may be either a shift matching, a scale matching or a combination of both the shift matching and the scale matching.

6. The method of claim 5 , wherein the shift matching is one of either a four-directional matching or an eight-directional matching, and the scale matching is one of either an enlarged matching or a reduced matching.

7. The method of claim 4 , wherein the motion matching method comprises:

determining a highest probability of an object movement in consideration of an object movement in a previous input image;

template matching in an order from a type of matching with the highest to lowest probability of having an object movement; and

terminating the template matching upon obtaining a result of a template matching which meets a predetermined threshold value.

8. The method of claim 4 , wherein in the step of detecting an object region from an input image by a template matching based upon an effective range, the speed estimation matching method comprises:

predicting a shift of an object region during the defined time interval (Δt) between an n-th input image and an input image in which an object region has been detected;

skipping the input frames between the n-th input image and the input image in which an object region has been detected; and

detecting an object region at a shifted position in the n-th input image after a time period of Δtxn.

9. A method of detecting a specific object in an image signal comprising:

dividing a full color range of the specific object into a plurality of sections, where each section includes a portion of the full color range;

obtaining a color image by detecting colors of the image which satisfy a limited color range of at least one section; and

detecting the object from the obtained color image, wherein obtaining a color image by detecting colors of the image which satisfy a limited color range of at least one section comprises obtaining color images by defining a window which includes at least one section, shifting the window by a predetermined distance, and detecting colors of the image which satisfy a color range limited by each shifting of the window.

10. The method of claim 9 , further comprising confirming the detected object using a general detection algorithm.

11. The method of claim 9 , wherein dividing a full color range of the specific object into a plurality of sections comprises determining the limited color range of each section by a statistical range which represents a size of the color range of the object to be detected in an image.

12. The method of claim 9 , wherein dividing a full color range of the specific object into a plurality of sections comprises dividing the full color range into the plurality of sections based upon a brightness component.

13. The method of claim 9 , wherein shifting the window from a present position to a next position by the predetermined distance equivalent to half the window size.

14. The method of claim 9 , wherein detecting the object from the obtained color image comprises detecting the object in an order from a color image corresponding to a limited color range with a highest to lowest probability of including the color of the object.

15. A method of detecting a specific object in an image signal, comprising:

judging whether a current input image frame is a first image frame;

detecting an initial object region from the current input image frame using a general detection algorithm if the current input image frame is the first input image frame; and

detecting an object region from an input image frame by a template matching based upon an effective range if the current input image frame is not the first input image frame.

16. The method of claim 15 , wherein detecting an initial object region from the current input image frame using a general detection algorithm comprises detecting the initial object region by either a template matching method using the current input image frame, or a method using an image difference and color information.

17. The method of claim 15 , wherein detecting an object region from the current input image frame by a template matching based upon an effective range composes defining the effective range as an estimated range where an object region may appear in a next input image frame.

18. The method of claim 15 , wherein detecting an object region from the current input image frame by a template matching based upon an effective range comprises performing the template matching by at least one of a whole matching method based on template matchings in predetermined positions, a motion estimation matching method based on a motion of an object, and a speed estimation matching method based on a relative speed difference of object motion appearing in the input images during a defined time interval (Δt).

19. The method of claim 18 , wherein the template matching in the whole matching method may be either a shift matching, a scale matching, or a combination of both the shift matching and the scale matching.

20. The method of claim 19 , wherein the shift matching is one of either a four-directional matching or an eight-directional matching, and the scale matching is one of either an enlarged matching or a reduced matching.

21. The method of claim 18 , wherein the motion matching method comprises:

determining a highest probability of an object movement in consideration of an object movement in a previous input image frame;

template matching in an order from a type of matching with the highest to lowest probability of having an object movement; and

terminating the template matching upon obtaining a result of a template matching which meets a predetermined threshold value.

22. The method of claim 18 , wherein, in the step of detecting an object region from the current input image frame by a template matching based upon an effective range, the speed estimation matching method comprises:

predicting a shift of an object region during the defined time interval (Δt) between an n-th input image frame and an input image frame in which an object region has been detected;

skipping the input frames between the n-th input image and the input image frame in which an object region has been detected; and

detecting an object region at a shifted position in the n-th input image frame after a time period of Δtxn.

23. A method of detecting a specific object in an image signal, comprising:

detecting an initial object region from an input image using a general detection algorithm if an object region has not been detected in a previous input image; and

detecting an object region from an input image by a template matching based upon an effective range if an object region was detected in a previous input image, wherein detecting an object region from an input image by a template matching based upon an effective range comprises performing the template matching by at least one of a whole matching method based on template matchings in predetermined positions, a motion estimation matching method based on a motion of an object, and a speed estimation matching method based on a relative speed difference of object motion appearing in the input images during a defined time interval (Δt), and wherein the motion matching method comprises:

determining a highest probability of an object movement in consideration of an object movement in a previous input image;

template matching in an order from a type of matching with the highest to lowest probability of having an object movement; and

terminating the template matching upon obtaining a result of a template matching which meets a predetermined threshold value.

24. A method of detecting a specific object in an image signal, comprising:

detecting an initial object region from an input image using a general detection algorithm if an object region has not been detected in a previous input image; and

detecting an object region from an input image by a template matching based upon an effective range if an object region was detected in a previous input image, wherein detecting an object region from an input image by a template matching based upon an effective range comprises performing the template matching by at least one of a whole matching method based on template matchings in predetermined positions, a motion estimation matching method based on a motion of an object, and a speed estimation matching method based on a relative speed difference of object motion appearing in the input images during a defined time interval (Δt), and wherein the speed estimation matching method comprises:

predicting a shift of an object region during the defined time interval (Δt) between an n-th input image and an input image in which an object region has been detected;

skipping the input frames between the n-th input image and the input image in which an object region has been detected; and

detecting an object region at a shifted position in the n-th input image after a time period of Δtxn.