METHOD FOR IDENTIFYING AN OBJECT IN A VIDEO ARCHIVE

• US 20120039506A1
• Filed: 08/26/2009
• Published: 02/16/2012
• Est. Priority Date: 08/27/2008
• Status: Active Grant

• Current Assignee: European Aeronautic Defence and Space Company EADS France (Airbus SE)
• Original Assignee: European Aeronautic Defence and Space Company EADS France (Airbus SE)
• Inventors: Sturzel, Marc, Jurie, Frédéric
• Granted Patent: US 8,594,373 B2
• Time in Patent Office: Days
• US Class Current: 382/103
• CPC Class Codes: G06F 16/70 of video data
• CPC Class Codes: G06V 20/47 Detecting features for summarising video content
• CPC Class Codes: G06V 20/54 of traffic, e.g. cars on the road, trains or boats

Abstract

The invention concerns a method for identifying an object in a video archive including multiple images acquired in a network of cameras including a phase of characterisation of the object to be identified and a phase of searching for the said object in the said archive, where the said characterisation phase consists in defining for the said object at least one semantic characteristic capable of being extracted, even in low-resolution images, from the said video archive.

Claims

1. A method for identifying an object in a video archive comprising a plurality of images acquired in a network of cameras, the method comprising:

a phase of characterization of the object to be identified and a phase of searching for the object to be identified in the video archive,

wherein the characterization phase comprises defining for the object at least one semantic characteristic capable of being extracted from the video archive, even for low-resolution images, the at least one semantic characteristic being directly interpretable by an operator, and

wherein the search phase comprises;

filtering the images from the video archive according to the previously defined semantic characteristic;

extracting automatically from the archive the images containing an object having the semantic characteristic;

defining a group of objects comprising all the objects present in the video archive having the semantic characteristic; and

measuring the similarity of the object to be identified with every other object of the previously defined group according to the visual and space-time characteristics in relation to the itinerary of the object to be identified in the space covered by the network of cameras.

2. A method according to claim 1, wherein measuring the similarity comprises:

estimating the compatibility of the semantic characterization of the object to be identified with the semantic characteristic extracted from the images of the other objects in the previously defined group, and/or,

estimating the space-time compatibility of the itinerary of the object to be identified with the itinerary of another object in the previously defined group having a similar semantic characteristic, wherein the another object does not have any characteristic incompatible with that of the object to be identified.

3. A method according to claim 2, further comprising assigning a probability coefficient to each measurement of similarity.

4. A method according to claim 3, further comprising merging the results of the measurements of the similarity so as to define a single unified measurement of similarity;

and enabling a distance in the space of the objects to be identified to be defined.

5. A method according to claim 4, further comprising undertaking a dynamic structuring of the space of the objects to be identified by means of the distance so as to navigate interactively in the video archive using a hierarchical tree.

6. A method according to claim 5, wherein the object to be identified is a human being.

7. A method according to claim 6, wherein the semantic characteristic of the human being comprises a visible physical characteristic.

8. A method according to claim 6, wherein the semantic characteristic of the human being comprises a visible clothing accessory.

9. A device for identifying an object in a video archive comprising a plurality of images acquired in a network of cameras, the device comprising:

a characterization module configured to characterize the object to be identified and a search module configured to search for the object in the video archive, wherein the characterization modules comprises;

means for defining for the object at least one semantic characteristic capable of being extracted from the video archive, even for low-resolution images, the at least one semantic characteristic being directly interpretable by an operator, and

wherein the search module comprises;

means for filtering the images from the video archive according to the previously defined semantic characteristic;

means for extracting automatically from the archive the images containing an object having the semantic characteristic;

means for defining a group of objects comprising all the objects present in the video archive having the semantic characteristic; and

means for measuring the similarity of the object to be identified with every other object of the previously defined group according to the visual and space-time characteristics in relation to the itinerary of the object to be identified in the space covered by the network of cameras.

10. A device according to claim 9, wherein the means for measuring similarity comprises:

a first computation module configured to estimate the compatibility of the semantic characteristic of the object to be identified with the semantic characteristic extracted from the images of the other objects of the previously defined group; and

/or,

a second computation module configured to estimate the space-time compatibility of the itinerary of the object to be identified with the itinerary of another object in the previously defined group having a semantic characteristic similar to that of the object to be identified.

11. A computer program recorded on a non-transitory recording medium and suitable, when it is executed on a computer, for identifying an object in a video archive comprising a plurality of images acquired in a network of cameras, the computer program comprising:

instructions to execute a phase of characterization of an object to be identified and instructions to execute a phase of searching for the object to be identified in the video archive,

wherein the characterization phase comprises defining for the object at least one semantic characteristic capable of being extracted from the video archive, the at least one semantic characteristic being directly interpretable by an operator,

and wherein the search phase comprises filtering the images from the video archive according to the previously defined semantic characteristic;

extracting automatically from the video archive the images containing an object having the semantic characteristic;

defining a group of objects comprising all the objects present in the video archive having the semantic characteristic; and

measuring the similarity of the object to be identified with every other object of the previously defined group according to the visual characteristics and constraints on the space-time itinerary of the object to be identified in the space covered by the network of cameras.