Image Analysis

• US 20090252421A1
• Filed: 06/30/2006
• Published: 10/08/2009
• Est. Priority Date: 07/28/2005
• Status: Active Grant

• Current Assignee: British Telecommunications PLC (BT Group PLC)
• Original Assignee: British Telecommunications PLC (BT Group PLC)
• Inventors: Stentiford, Frederick W.M.
• Granted Patent: US 8,135,210 B2
• Time in Patent Office: Days
• US Class Current: 382/209
• CPC Class Codes: G06T 7/00 Image analysis
• CPC Class Codes: G06V 10/44 Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
• CPC Class Codes: G06V 20/00 Scenes; Scene-specific elements control of digital cameras H04N23/60

Abstract

An image, represented by an ordered set of elements (x_i) each having a value is analysed in order to detect vanishing points. The method comprises, for each of a plurality of root positions (x₀), repeatedly performing the steps of: (i) selecting a first plurality of elements (x_i) from the ordered set; (ii) for each selected element (x_i), selecting a second element (u_i) such that the selected second element has a vector position relative to the root position that is scaled by a factor (α) in comparison with the position of the first selected element; (iii) determining whether the selected elements meet a match criterion requiring that the value of each of the first elements is similar to the value of the corresponding second element; and (iv) in the event of a match, updating a similarity score (H) in respect of that root element. Once these scores have been found, they can be examined a part of the image corresponding to a peak value of the similarity score.

Claims

1. A method of analysing an image represented by an ordered set of elements (x_i) each having a value comprising, for each of a plurality of root positions (x₀), repeatedly performing the steps of:

(i) selecting a first plurality of elements (x_i) from the ordered set;

(ii) for each selected element (x_i), selecting a second element (u_i) such that the selected second element has a vector position relative to the root position that is scaled by a factor (α

) in comparison with the position of the first selected element;

(iii) determining whether the selected elements meet a match criterion requiring that the value of each of the first elements is similar to the value of the corresponding second element;

and (iv) in the event of a match, updating a similarity score (H) in respect of that root element.

2. A method according to claim 1 including the further step of examining the similarity scores (H) to identify a part of the image corresponding to a peak value of the similarity score.

3. A method according to claim 1 in which all the root position are positions of elements within the image.

4. A method according to claim 1 in which at least some of the root positions are positions outside the image.

5. A method according to claim 1 in which, in the step of selecting a first plurality of elements, the selection is constrained such that no more than a set number of the selected elements meet a predetermined criterion of similarity to others of the elements.

6. A method according to claim 1 in which the scaling factor (α

) is fixed.

7. A method according to claim 1 which the scaling factor (α

) is freshly chosen for each step of selecting second elements.

8. A method according to claim 1 in which the scaling factor (α

) is in the range 0.3 to 0.8.

9. A method according to claim 1 in which the images is a colour image, each value comprises brightness and/or colour components, and the match criterion is that the difference between each component of the respective first element and the corresponding component of the corresponding second element does not exceed a set amount.

10. A method of analysing an image represented by an ordered set of elements (x_i) each having a value comprising, for each of a plurality of reference directions, repeatedly performing the steps of:

(i) selecting a first plurality of elements (x_i) from the ordered set;

(ii) for each selected element (x_i), selecting a second element (u_i) such that the selected second element has a vector position relative to the root position that is, in comparison with the position of the first selected element, shifted in the respective reference direction by a shift amount (Δ

);

(iii) determining whether the selected elements meet a match criterion requiring that the value of each of the first elements is similar to the value of the corresponding second element; and

(iv) in the event of a match, updating a similarity score (H) in respect of that reference direction.