Baggage identification method

Patent metadata

Field	Value
Publication number	US-11049234-B2
Application number	US-202016809151-A
Country	US
Kind code	B2
Filing date	Mar 4, 2020
Priority date	Mar 22, 2019
Publication date	Jun 29, 2021
Grant date	Jun 29, 2021

Abstract

A method identifies a non-identified baggage item based on: acquisition of at least two images of the non-identified baggage of different types, the type of one image of a baggage item dependant on a point-of-view on the baggage in the image and/or on the nature of the data representing the baggage in the image taken among different natures of data comprising data representing the visible range and/or infrared data and/or data representing three-dimensional information; and, for each image type, on use of a neural network suited to the type for classifying the baggage represented by the image in a class of baggage classes defined for the type. Once obtained, the classes allow seeking of baggage corresponding to each class identified in a baggage database. Each baggage item in the database is associated with the baggage addressee. Each corresponding baggage item is compared with the non-identified baggage to identify the addressee.

First claim

The invention claimed is:

1. A method for identifying a baggage item devoid of any identifier representing an addressee of said baggage, referred to as non-identified baggage, comprising:
obtaining at least two images of the non-identified baggage of different types, the type of image of a baggage item being dependent on a point of view on said baggage represented by said image and the nature of the data representing said baggage in said image taken from a plurality of natures of different data comprising data representing the visible range and/or data representing the infrared range and/or data representing three-dimensional information;
for each image of the non-identified baggage obtained:
applying said image to a convolutional neural network trained on baggage images of the same type as said image in order to obtain, for each class in a plurality of baggage classes defined for the type of image, a probability of said non-identified baggage belonging to said class; and
identifying the baggage classes having the highest probability;

applying a baggage search procedure comprising: searching for baggage corresponding to each class identified in a baggage database in order to obtain at least one candidate baggage item, each baggage item in said database being associated with an identifier representing an addressee of the baggage; and
comparing the non-identified baggage with each candidate baggage item in order to identify the addressee of said non-identified baggage.

2. The method according to claim 1 , wherein the method further comprises: obtaining at least one item of information representing the non-identified baggage among a weight, a colour and at least one dimension, the baggage search procedure further comprising seeking in the database baggage corresponding to each item of information representing said non-identified baggage.

3. The method according to claim 1 , wherein each baggage item referenced in the database is associated with timestamping information representing a time of checking in of said baggage, the baggage search procedure further comprising seeking in the database baggage associated with timestamping information corresponding to a period determined from a time of discovery of the non-identified baggage.

4. The method according to claim 1 , wherein each baggage item referenced in the database is associated with a number of items of classification information equal to a number of types of image considered at the time of each classification of a baggage item in the database, each type of image used for the non-identified baggage corresponding to at least one type of image considered at the time of each classification of a baggage item in the database, each item of classification information corresponding to a type of image and representing a class in a plurality of classes associated with said type in which the baggage was classified.

5. A device for identifying a baggage item devoid of any identifier representing an addressee of said baggage, referred to as non-identified baggage; comprising at least one processor configured to:
obtain at least two images of the non-identified baggage, of different types, the type of image of a baggage item being dependent on a point of view on said baggage represented by said image and the nature of the data representing said baggage in said image taken from a plurality of natures of different data comprising data representing the visible range and/or data representing the infrared range and/or data representing three-dimensional information;
for each image of the non-identified baggage obtained:
apply said image to a convolutional neural network trained on images of baggage of the same type as said image in order to obtain, for each class in a plurality of classes of baggage defined for the type of image, a probability of said non-identified baggage belonging to said class; and
identify the class of baggage having the highest probability;

apply a baggage search procedure comprising: searching for baggage corresponding to each class identified in a baggage database in order to obtain at least one candidate item of baggage; each item of baggage in said database being associated with an identifier representing an addressee of the baggage; and
compare the non-identified baggage with each candidate item of baggage in order to identify the addressee of the non-identified baggage.

6. A non-transitory storage medium storing a computer program comprising instructions for the implementation, by a device, of the method according to claim 1 , when said program is executed by a processor of said device.

Assignees

• Idemia Identity & Security France

Inventors

• Maurin Franck
• Barre Anthony

Classifications

• G06V10/764 Primary

  using classification, e.g. of video objects · CPC title

• G06V20/64

  Three-dimensional [3D] objects · CPC title

• G06T7/0002 Primary

  Inspection of images, e.g. flaw detection · CPC title

• G06F18/2431

  Multiple classes · CPC title

• G06F18/2415

  based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate · CPC title