CHIESA Valeria

Thesis

Soft Biometrics

The subject of Valeria Chiesa thesis is related to soft biometrics and image processing. Her thesis is partly financed by the European projects IDEA4SWIFT and IDENTITY.

 

Soft biometrics

Classical biometry offers a natural and reliable solution for establishing the identity of an individual. The use of human physical and behavioral characteristics has been increasingly adopted in security applications due to various advantages, such as universality, robustness, permanence and accessibility. Currently state-of-the-art intrusion detection and security mechanism systems include meanwhile by default at least one biometric trait. The latest addition of soft biometry inherits a main part of the advantages of classical biometry and furthermore endorses by its own assets.

The beginnings of soft biometrics science were laid by Alphonse Bertillon in the 19th century, who firstly introduced the idea for a personal identification system based on biometric, morphological and anthropometric determinations. He used traits like colors of eye, hair, beard and skin; shape and size of the head; general discriminators like height or weight and also description of indelible marks such as birth marks, scars or tattoos. A great majority of those descriptors fall at the present time into the category of soft biometrics. Jain et al. first introduced the term soft biometrics to be a set of characteristics that provide some information about the individual, but are not able to individually authenticate the person, mainly due to the lack of distinctiveness and permanence. Later on, it was also noted that soft biometrics are not expensive to compute, can be sensed at a distance, do not require the cooperation of the surveillance subjects and have the aim to narrow down the search from a group of candidate individuals. Moreover we here note that the human compliance of soft biometrics is a main factor, which differentiates soft biometrics from classical biometrics offering new application fields.

Nowadays it is possible to define the soft biometric traits as physical, behavioral or adhered human characteristics, classifiable in pre-defined human compliant categories. These categories are, unlike in the classical biometric case, established and time-proven by humans with the aim of differentiating individuals. In other words the soft biometric traits instances are created in a natural way, used by humans to distinguish their peers. We note that the human compliant labeling is sometimes also referred to as semantic annotation.

 

 

 

 

IDEA4SWIFT

Identity, Documents, Exchange Authentication 4 Systems Worldwide Interconnections of Frequent Travelers

The border has a singular position in our society. It is either a fence built to protect citizens against external threats or a place of exchange and meeting. Improving the security of border management in Europe through intelligence and facilitating the legitimate free movement of people and goods is an issue to consider. The EU emphasizes automation and Automatic Border Gate experiences have already been launched in the main airports (Frankfurt, Schiphol, CDG...) and  this trend is becoming a global one. The IDEA4SWIFT project will support this effort to extend the population of targeted users while existing standalone systems can be used by a very limited number of users. Going beyond those limits is a real technological challenge with the integration of the latest improvements in security (e.g. passport verification). The project aims to design, develop and integrate disruptive technologies in biometry and soft-biometry in a real environment, envisages strong enhancement for authentication protocols for citizens crossing frontiers and secure communications for sensitive systems.

IDEA4SWIFT project involves academic institutions as well as companies from France and Turkey. Gemalto, Morpho, EURECOM, Institut Mines-Télécom, id3 Semiconducteurs, Oberthur Technologies, Proline, Smartsoft and Mozaik work together in order to achieve the aimed goal.

 

First year

Within the context of IDEA4SWIFT, the work will consist in extracting automatically some information from face images like gender and age. The goal is then to check if these information are fully compliant with information available/readable on passports.

Age estimation is still a challenging problem in soft biometrics because the face aging process is influenced by a large variety of factor, both intrinsic, like genetic factor, or extrinsic, as the environment. To estimate a precise age has been revealed an hard target; for this reason most of the studies are focused on allocation of subjects in a defined class, as childhood, adolescence, youth, adult age or old age. For example in (M. Z. Lazarus K. S., 2013) authors work on synthesizing images of human faces analyzing wrinkles in order to estimate the age. In (Hu Han, 2013) a biologically inspired model is used to extract features able to train a SVM. In (Anuradha Yarlagadda, 2015) Correlation Fractal Dimension of facial edges is explored. Additional knowledges can be useful in age estimation: for this reason in (Ramesha K., 2010) authors choose to combine gender recognition with age classification.
In the specific context of IDEA4SWIFT, a mismatch between the apparent age of the traveler and the passport information can be a good alarm in order to identify fake documents. With this aim, a tool enables to classify two face images in terms of dates will be designed. The goal is to define from two different pictures of a specific person in which the person is younger/older. Until now classical approaches obtain an error around 7%, according to the database used. The idea is to compare these techniques based on wrinkles detection with more innovative methods based on deep learning. 

Another domain of interest concerns new video sensors and in particular plenoptic cameras. This kind of technology allows to capture four-dimensional information about the light rays in a scene in a single exposure. The objective here is to specify what could be the advantages to use new sensors in face recognition, and to design new adhoc algorithms to process such data. This study will start by the compilation of the art-of-the-art in plenoptic imaging, as such a document does not exist yet.

Second year

Within the context of IDEA4SWIFT, the work will consist in detecting possible spoofing attacks and counter-measures  The main objective is to make the distinction between a real face and a photography displayed in front of a video camera. The benefits of using new sensors (e.g. plenoptic camera) in this field have already been proved in (Sooyeon Kim, 2014) and (R. Raghavendra, 2015). A deeper study of light field images can improve an anti-spoofing method.

The second year may include a visiting stay in one of the extra European partners of the Marie Curie Mobility Project IDENTITY.

Third year

The content of the third year will be described in one year.