Object-end positioning method and system

• US 8,396,300 B2
• Filed: 05/11/2009
• Issued: 03/12/2013
• Est. Priority Date: 12/08/2008
• Status: Active Grant

• Current Assignee: Industrial Technology Research Institute
• Original Assignee: Industrial Technology Research Institute
• Inventors: Li, Jia-Chang, Kuo, Chien-Chun, Wang, Ko-Shyang, Chen, Po-Lung
• Primary Examiner(s): Wang, Claire X
• Assistant Examiner(s): ROWLAND, JHERE M
• Application Number:
  US12/463,508
• Publication Number:
  US 20100142831A1
• Time in Patent Office: 1,401 Days
• Field of Search: None
• US Class Current: 382/216
• CPC Class Codes: G06V 10/46 Descriptors for shape, contour or point-related descriptors, e.g. scale invariant feature transform [SIFT] or bags of words [BoW]; Salient regional features colour feature extraction G06V10/56
• CPC Class Codes: G06V 40/10 Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands

Abstract

An object-end positioning method is provided for positioning lower ends of two limbs of an object. In the method, a foreground processing is performed on an original image to obtain a foreground image. A number of turning points are obtained according to the foreground image, wherein connection of the turning points forms a polygonal curve. Each turning point is classified to be a convex or concave point according to an included angle between lines connecting the turning point to two adjacent turning points. A number of selected convex points and selected concave points are selected. Two of the selected convex points are selected as two temporary ends. Connection of the two temporary ends and a selected concave point located between the two temporary ends forms a triangle. Two positioning ends for positioning the lower ends of the two limbs of the object are determined according to the two temporary ends.

Claims

1. An object-end positioning method, for positioning lower ends of two limbs of an object, the method comprising:

obtaining an original image corresponding to an image information of the object;

performing a foreground processing on the original image to obtain a foreground image corresponding to a contour of the object;

obtaining a plurality of turning points according to the foreground image, wherein the turning points are positioned along the contour and connection of the turning points with straight line segments forms a single polygonal curve substantially approximate to the contour of the object in the original image;

determining each of the turning points to be a convex point or a concave point, and selecting a plurality of selected convex points and a plurality of selected concave points along a predetermined direction, wherein the convex point and two straight line segments adjacent to the convex point form an interior angle of the single polygonal curve less than 180 degrees, and the concave point and two straight line segments adjacent to the concave point form an interior angle of the single polygonal curve larger than 180 degrees;

selecting two of the selected convex points as two determined ends, wherein connection of the two determined ends and a selected concave point located between the two determined ends forms a triangle corresponding to the two limbs of the object in the original image; and

determining two positioning ends for positioning the lower ends of the two limbs of the object according to the two determined ends;

wherein the two determined ends indicate the lower ends of the two limbs of the object, respectively.

2. The method according to claim 1, wherein the predetermined direction is related to a direction oriented from the lower end to the upper end in each of the two limbs of the object in the original image.

3. The method according to claim 1, wherein before the step of selecting two of the selected convex points as the two determined ends, the method further comprises:

judging if the selected convex points and the selected concave points meet a triangular-characteristic match; and

if the two of the selected convex points and one of the selected concave points meet the triangular-characteristic match, determining the two of the selected convex points as the two determined ends.

4. The method according to claim 3, wherein the step of determining the two of the selected convex points as the two determined ends further comprises:

if the two of the selected convex points and one of the selected concave points meet the triangular-characteristic match, and another two of the selected convex points and another one of the selected concave points also meet the triangular-characteristic match, judging if an area formed by the two of the selected convex points and the one of the selected concave points is larger than an area formed by the another two of the selected convex points and the another one of the selected concave points, and if yes, determining the two of the selected convex points as the two determined ends.

5. The method according to claim 3, wherein the step of determining if the selected convex points and the selected concave points meet the triangular-characteristic match comprises:

judging if a slope of a connection line between the two of the selected convex points is smaller than a predetermined slope according to a vector perpendicular to the predetermined direction, and judging if a position of the one of the selected concave points projected to the vector is located between positions of the two of the selected convex points projected to the vector.

6. The method according to claim 3, wherein if a side close to the lower ends of the two limbs of the object is defined to be a lower side, the predetermined direction is a direction from the lower side to an upper side, and the step of determining the two of the selected convex points as the two determined ends comprises:

judging if the two of the selected convex points is closer to the lower side than the one of the selected concave point, and if yes, determining the two of the selected convex points as the two determined ends.

7. The method according to claim 3, wherein after the step of judging if the selected convex points and the selected concave points meet the triangular-characteristic match, the method further comprises:

if the selected convex points and the selected concave points do not meet the triangular-characteristic match, obtaining two tracking ends according to two previous ends, wherein the two previous ends are lower ends of the two limbs of the object positioned from a previous original image;

wherein the step of determining the two positioning ends further comprises determining the two positioning ends according to the two tracking ends.

8. The method according to claim 7, wherein the step of obtaining the two tracking ends further comprises:

determining the two tracking ends according to a brightness variation of surrounding pixels of the two previous ends in the previous original image and the original image.

9. The method according to claim 7, wherein the step of obtaining the two tracking ends further comprises:

determining the two tracking ends according to a color variation of surrounding pixels of the two previous ends in the previous original image and the original image.

10. The method according to claim 7, wherein the step of obtaining the two tracking ends further comprises:

determining the two tracking ends by a prediction or probability method according to positions of the two previous ends and another two previous ends;

wherein the another two previous ends are lower ends of the two limbs of the object positioned from another one previous original image, and the original image, the previous original image and the another one previous original image are continuously obtained.

11. An object-end positioning system for positioning lower ends of two limbs of an object, the system comprising:

a capturing circuit, for obtaining an original image corresponding to an image information of the object;

a processor, for performing a foreground processing on the original image to obtain a foreground image corresponding to a contour of the object, wherein the processor obtains a plurality of turning points according to the foreground image, wherein the turning points are positioned along the contour and connection of the turning points with straight line segments forms a single polygonal curve substantially approximate to the contour of the object in the original image, and the processor determines each of the turning points to be a convex point or a concave point and selects a plurality of selected convex points and a plurality of selected concave points along a predetermined direction, wherein the convex point and two straight line segments adjacent to the convex point form an interior angle of the single polygonal curve less than 180 degrees, and the concave point and two straight line segments adjacent to the concave point form an interior angle of the single polygonal curve larger than 180 degrees;

a matching circuit, for selecting two of the selected convex points as two determined ends, wherein connection of the two determined ends and a selected concave point located between the two determined ends forms a triangle corresponding to the two limbs of the object in the original image; and

a positioning circuit, for determining two positioning ends for positioning the lower ends of the two limbs of the object according to the two determined ends;

wherein the two determined ends indicate the lower ends of the two limbs of the object, respectively.

12. The system according to claim 11, wherein the predetermined direction is related to a direction oriented from the lower end to the upper end in each of the two limbs of the object in the original image.

13. The system according to claim 11, wherein when the matching circuit selects the two of the selected convex points as the two determined ends, the matching circuit judges if the selected convex points and the selected concave points meet a triangular-characteristic match, and if the matching circuit determines that the two of the selected convex points and one of the selected concave points meet the triangular-characteristic match, the matching circuit determines to select the two of the selected convex points as the two determined ends.

14. The system according to claim 13, wherein if the matching circuit determines that the two of the selected convex points and the one of the concave points meet the triangular-characteristic match and another two of the selected convex points and another one of the selected concave points also meet the triangular-characteristic match, the matching circuit further judges if an area formed by the two of the selected convex points and the one of the selected concave point is larger than an area formed by the another two of the selected convex points and the another one of the selected concave points, and if yes, the matching circuit determines the two of the selected convex points as the two determined ends.

15. The system according to claim 13, wherein when the matching circuit judges if the selected convex points and the selected concave points meet the triangular-characteristic match, the matching circuit judges if a slope of a connection line between the two of the selected convex points is smaller than a predetermined slope according to a vector perpendicular to the predetermined direction, and judges if a position of the one of the selected concave points projected to the vector is located between two positions of the two of the selected convex points projected to the vector.

16. The system according to claim 13, wherein if a side close to the lower ends of the two limbs of the object is defined to be a lower side, the predetermined direction is a direction from the lower side to an upper side, and the two of the selected convex points are closer to the lower side than the one of the selected concave points.

17. The system according to claim 13, wherein the matching circuit further comprises:

a tracking circuit, for obtaining two tracking ends according to two previous ends, wherein the two previous ends are lower ends of the two limbs of the object positioned from a previous original image;

wherein when the selected convex points and the selected concave points do not meet the triangular-characteristic match, the positioning circuit determines the two positioning ends according to the two tracking ends.

18. The system according to claim 17, wherein when the tracking circuit obtains the two tracking ends, the tracking circuit determines the two tracking ends according to a brightness variation of surrounding pixels of the two previous ends in the previous original image and the original image.

19. The system according to claim 17, wherein when the tracking circuit obtains the two tracking ends, the tracking circuit determines the two tracking ends according to a color variation of surrounding pixels of the two previous ends in the previous original image and the original image.

20. The system according to claim 17, wherein when the tracking circuit obtains the two tracking ends, the tracking circuit determines the two tracking ends by a prediction or probability method according to positions of the two previous ends and another two previous ends;

wherein the another two previous ends are lower ends of the two limbs of the object positioned from another one previous original image by the positioning circuit, and the original image, the previous original image and the another one previous original image are continuously obtained by the capturing circuit.