Idemia Identity & Security France

Fantin Girard

Cédric Thuillier

Emilie Niaf

Guy Mabyalaht

Cedric Thuillier

Jonathan Milgram

Emine Krichen

A method for matching an image of a palm of an individual with at least one image from a plurality of reference palm images, which includes: creation of a reference base from the plurality of reference palm images comprising, for each reference palm image, a selection of at least one reference palm region and a coding into a reference regional biometric information item, determination of an individual palmary biometric information item from the image of the palm of the individual comprising a selection of at least one individual palm region and the coding thereof into individual regional biometric information item, for each image, comparison of each individual regional biometric information item with each reference regional biometric information item, and matching of the image of the palm of the individual with at least one image from the plurality of reference palm images, as a function of a result of the comparison.

Method for matching an image of a palm of an individual with at least one image from a plurality of reference palm images

Methods for augmenting a training image base representing a print on a background, for training parameters of a convolutional neural network, CNN, or for classification of an input image The present invention relates to a method for augmenting a training image base representing a print on a background, characterized in that it comprises the implementation, by data processing means () of a server (), of steps of:

Method for augmenting a training image base representing a print on a background by means of a generative adversarial network

A parameter training method for a convolutional neural network (CNN) for classifying image type data representative of a biometric trait. The method includes the implementation, by a data processor of a server, the steps of (a) for at least one data item from an already classified training database, generation of several alternate versions of this data each by application of at least one transformation chosen from a set of reference transformations satisfying a statistical distribution of transformations observed in the training database and (b) training the parameters of the CNN, from the already classified training database augmented with said alternate versions of the data.

Parameter training method for a convolutional neural network

The present invention concerns a method of classification of an input image representative of a biometric trait by means of a first convolutional neural network, CNN, characterized in that it comprises the implementation by data processing means () of a client () of steps of:

Method of classification of an input image representative of a biometric trait by means of a convolutional neural network

The present invention relates to a method for extracting a feature vector from an input image representative of an iris by means of a neural network, characterized in that it can be trained end-to-end and comprises the implementation by data-processing means of a client of steps of:

Method for extracting a feature vector from an input image representative of an iris by means of an end-to-end trainable neural network

A method for reconstructing an imprint image, from a set of image portions, includes the steps of: extracting, from each image portion, a set of local points of interest and, for each point of local interest, calculating a descriptor vector that characterizes said point of local interest; for each pair of two image portions, evaluating a local interest points association score representative of a probability that the two image portions are contiguous on the imprint image; assembling the image portions of a best pair to form an assembled fragment; repeating the above steps by replacing each time, in the set of image portions, the two image portions of the best pair, until all the association scores of the remaining pairs are less than or equal to a predetermined threshold, and producing an assembly map of the image portions; merging the image portions to reproduce the imprint image.

Growing community of inventors

Courbevoie, France

Fantin Girard