US 9,866,894 B2Granted

Method for annotating an object in a multimedia asset

AssigneeAlcatel Lucent

InventorsSammy Lievens, Donny Tytgat

Life of this patent

Estimating the term…

Loading patent…

Abstract

The present invention relates to a method, system and related devices for annotating an object in a multimedia asset, where the multimedia asset comprises at least one image. The method comprises the steps of first determining a category of the object in an image of the multimedia asset by means of classifying the object and subsequently selecting based on the category of the object an appropriate object detector model from at least one object detector model associated with said category. Thereafter a location of the object in the image of the multimedia asset is determined based on the object detector model that is selected, where the object detector model corresponds to the object which step is followed by segmenting the object detected in the image of the multimedia asset based on the location of the object within the said image and optionally based on the object detector. Finally, the object within the image being segmented is annotated with a hyperlink.

Claims9

Dependent claims

A method for annotating an object in a multimedia asset, said multimedia asset comprising at least one image, said method comprising:

determining a category of said object in an image of said multimedia asset based on classifying said object by extracting features from said image and performing a statistical analysis on said extracted features to identify one or more object classes associated with said object;
selecting, based on said category of said object, an object detector model from a plurality of object detector models associated with said category of said object, said selected object detector model describing specific characteristics of said object;
determining a location of said object within said image of said multimedia asset, said location of said object being a region within said image in which said object is detected based on said selected object detector model corresponding to said object;
segmenting said object detected in said image of said multimedia asset, based on (i) said location of said object within said image and (ii) said selected object detector model, by delineating a shape of said object and marking a boundary of said object, said marked boundary enabling said object to be visually separated from background and other objects within said image;
and
automatically annotating said segmented object in said image (i) with a hyperlink, and (ii) by adding a condition for visual presentation of said segmented object, based on said marked boundary of said object within said image.

3 more elements

A multimedia annotating device for annotating an object in a multimedia asset, said multimedia asset comprising at least one image, said multimedia annotating device comprising:

a memory storing computer-readable instructions;
and
a processor configured to execute the computer-readable instructions to,
determine a category of said object in an image of said multimedia asset based on classifying said object by extracting features from said image and performing a statistical analysis on said extracted features to identify one or more object classes associated with said object;
select, based on said category of said object, an object detector model from a plurality of object detector models associated with said category of said object, said selected object detector model describing specific characteristics of said object;
determine a location of said object within said image of said multimedia asset, said location of said object being a region within said image in which said object is detected based on said selected object detector model corresponding to said object;
segment said object detected in said image of said multimedia asset, based on (i) said location of said object within said image and (ii) said selected object detector model, by delineating a shape of said object and marking a boundary of said object, said marked boundary enabling said object to be visually separated from background and other objects within said image;
and
automatically annotate said segmented object in said image (i) with a hyperlink, and (ii) by adding a condition for visual presentation of said segmented object, based on said marked boundary of said object within said image.

6 more elements

Dependencies are inferred from the claim text.

Facts

Application: 15/105,868Filed 11 Dec 2014
Priority: EP 1330674417 Dec 2013
Published: US 2016/0323627 A13 Nov 2016
Granted: 9 Jan 2018
Term ends: Not estimatedThe source records don’t support an estimate.

Classification10 codes

G06V 20/41
In video content: higher-level, semantic clustering, classification or understanding of video scenes, e.g. detection, labelling or Markovian…
G06F 18/285
Analysing: selection of pattern recognition techniques, e.g. of classifiers in a multi-classifier system
G06V 10/87
Using pattern recognition or machine learning: using selection of the recognition techniques, e.g. of a classifier in a multiple classifier system
H04N 21/431
Processing of content or additional data, e.g. demultiplexing additional data from a digital video stream; Elementary client operations, e.g.…:…