Motion characterisation

• US 7,860,272 B2
• Filed: 03/21/2007
• Issued: 12/28/2010
• Est. Priority Date: 03/23/2006
• Status: Expired due to Fees

• Current Assignee: Canon Information Systems Research Australia Pty Limited (Canon Inc.)
• Original Assignee: Canon Information Systems Research Australia Pty Limited (Canon Inc.)
• Inventors: Kerr, Julian Lewis
• Primary Examiner(s): Strege; John B
• Application Number:
  US11/689,269
• Publication Number:
  US 20070223808A1
• Time in Patent Office: 1,378 Days
• Field of Search: 382/103, 382/107, 382/155, 706/12, 706/14, 706/16, 706/20, 700/47, 700/48
• US Class Current: 382/103
• CPC Class Codes: G06F 18/40 Software arrangements specially adapted for pattern recognition, e.g. user interfaces or toolboxes therefor
• CPC Class Codes: G06V 10/945 User interactive design; Environments; Toolboxes
• CPC Class Codes: G06V 20/52 Surveillance or monitoring of activities, e.g. for recognising suspicious objects recognising microscopic objects G06V20/69

Abstract

A method for configuring a machine learning model for use in a motion characterisation process. The method comprises, in a processing system, acquiring user characterisations for respective portions of at least one video sequence and configuring the model using the user characterisations and at least one property associated with the respective portions. Any inconsistency in the user characterisations is then determined, using the model, with an indication of the inconsistency being displayed. This allows the user to select at least one option for addressing the inconsistency, with the model being reconfigured based on the selected at least one option.

Claims

1. A method for configuring a model for use in a motion characterisation process, wherein the method comprises, in a processing system:

a) acquiring user characterisations for respective portions of at least one video sequence;

b) configuring the model using the user characterisations and at least one property associated with the respective portions;

c) determining an inconsistency in the user characterisations using the model;

d) displaying an indication of the inconsistency;

e) determining a selected at least one option for addressing the inconsistency in accordance with user input commands; and

,

f) reconfiguring the model based on the selected at least one option.

2. A method according to claim 1, wherein the portions comprise spatiotemporal regions in frame sequences in the at least one video sequence.

3. A method according to claim 1, wherein the method comprises, in the processing system:

a) causing at least a portion of the at least one video sequence to be displayed to a user;

b) determining a region of the at least one video sequence designated by the user;

c) determining an adjustment factor designated by the user; and

,

d) determining the user characterisations using the user designated region and the adjustment factor.

4. A method according to claim 1, wherein the method comprises, in the processing system:

a) configuring the model using the user characterisations;

b) determining a characterisation for at least one region within at least one user designated region using the model;

c) comparing the determined characterisation to the user characterisations for the at least one user designated region; and

,

d) determining an inconsistency based on the result of the comparison.

5. A method according to claim 1, wherein the method comprises, in the processing system:

a) causing an indication of one or more options for addressing the inconsistency to be displayed; and

,

b) selecting at least one of the options in accordance with user input commands.

6. A method according to claim 1, wherein the method comprises;

in the processing system, causing the characterisation of a region to be displayed by at least one of;

a) indicating an extent of a characterised region using a polygon; and

,

b) indicating the characterisation using at least one of;

i) a colour associated with a polygon indicating a region extent; and

,

ii) a label.

7. A method according to claims 1, wherein the method comprises, in the processing system:

a) causing at least one portion of the at least one video sequence to be displayed to a user;

b) for one or more regions of the at least one video sequence, displaying a characterisation; and

,

c) determining a user characterisation in response to user input commands.

8. A method according to claim 1, wherein the method comprises;

in the processing system;

a) causing an adjustment indicator to be displayed together with at least one portion of the at least one video sequence; and

,

b) determining a user characterisation at least in part based on an adjustment factor determined based on designation of the adjustment indicator by the user.

9. A method according to claim 1, wherein the method comprises, in the processing system, determining at least one of the size, shape, position and user characterisation of a user designated region in accordance with user input commands.

10. A method according to claim 1, wherein the method comprises, in the processing system, tracking movement of a user designated region across a number of frames in the at least one video sequence in accordance with user input commands.

11. A method according to claim 1, wherein the method comprises, in the processing system, causing an indication of the inconsistency to be displayed by:

a) determining a number of instances in which user characterisation of motion is not correctly characterised using the configured model; and

,

b) causing the number to be displayed.

12. A method according to claim 1, wherein the method comprises, in the processing system, determining an inconsistency if the model does not correctly characterise motion characterised by a user characterisation.

13. A method according to claim 12, wherein the method comprises, in the processing system, addressing an inconsistency using at least one of:

a) a further user characterisation;

b) cancellation of a configuration of the model; and

,

c) an indication of an adjustment factor.

14. A method according to claim 13, wherein the model is used to characterise video sequences, and wherein user characterisations are used to modify the motion characteriser to thereby more accurately characterise video sequences.

15. A method according to claim 14, wherein the method comprises, in the processing system, using the user characterisations to adjust a motion detection sensitivity.

16. A method according to claim 15, wherein the method comprises, in the processing system:

a) for at least one region of at least one video sequence, determining a motion thread;

b) using the motion thread to determine a feature vector; and

,

c) characterising at least a portion of at least one video sequence using the feature vector.

17. A method according to claim 16, wherein the method further comprises, in the processing system:

a) determining a value of a sensitivity parameter; and

,

b) utilising the value of the sensitivity parameter to configure a motion characteriser.

18. A method according to claim 17, wherein the respective portions are formed from at least one of:

a) respective segments of at least one video sequence;

b) respective regions within a segment of at least one video sequence;

c) respective segments of different video sequences; and

,

d) respective regions within segments of different video sequences.

19. A method according to claim 1, wherein the method includes, in the processing system, displaying the inconsistency using a timeline.

20. A method according to claim 19, wherein the method includes, in the processing system:

a) displaying the timeline representing at least one video sequence; and

,

b) displaying at least one indicator on the timeline, the indicator representing the temporal location of an inconsistency within the at least one video sequence.

21. Apparatus for configuring a model for use in a motion characterisation process, wherein the apparatus comprises a processing system for:

a) acquiring user characterisations for respective regions of at least one video sequence;

b) configuring the model using the user characterisations and at least one property associated with the respective portions;

c) determining an inconsistency in the user characterisations using the model;

d) displaying an indication of the inconsistency;

e) determining a selected at least one option for fixing the inconsistency in accordance with user input commands; and

,

f) reconfiguring the model based on the selected at least one option.

22. A non-transitory computer-readable medium for configuring a model for use in a motion characterisation process, the non-transitory computer-readable medium including computer executable code which when executed on a suitable processing system causes the processing system to:

a) acquire user characterisations for respective regions of at least one video sequence;

b) configure the model using the user characterisations and at least one property associated with the respective regions;

c) determine an inconsistency in the user characterisations using the model;

d) display an indication of the inconsistency;

e) determine a selected at least one option for fixing the inconsistency in accordance with user input commands; and

,

f) reconfigure the model based on the selected at least one option.