| Titre : | Deployment of a Face Recognition System within a practical context | | Type de document : | projet fin études | | Auteurs : | Abdelkarim AZOUZI, Auteur | | Langues : | Français (fre) | | Catégories : | e-Management et Business Intelligence
| | Index. dĂ©cimale : | 2119/19 | | RĂ©sumĂ© : | The human face plays an important role in our social interaction, conveying people’s
identity. Using the human face as a key to security, face recognition technology
has received significant attention in the past several years due to its potential for a
wide variety of real-life applications.
The goal of this project is to study this technology and build a robust Face Recognition
system that can be compared to Human-Level performance. To achieve this
goal, we first had to go into the background underlying this technology, by searching
for and analyzing state-of-the-art algorithms. Hence, a face recognition system is
always a combination of two main concepts, face detection and face recognition, as
a consequence this project is tackling both concepts.
A survey of available machine learning algorithms will be presented by going
through the analysis of recently published benchmarks. Based on these benchmarks
results, a list of potential algorithms is selected, as well as two evaluation data-sets.
This selection allows to structure our own experiments, by taking those algorithms
and testing them with different evaluation protocols, and determine which algorithms
perform better and find out relevant algorithms for our system, that is to
say MTCNN [1] for detection and FaceNet [2] for recognition.
Finally, to have more reliable results with regards to this project, A prototyping
use case has been built in shape of biometric solution for payment. This work
describes all the steps required to design the use case. Then, we report on the development
details of our system and also its deployment process into the prototype.
The characterization of hardware and the description of software is presented. |
Deployment of a Face Recognition System within a practical context [projet fin études] / Abdelkarim AZOUZI, Auteur . - [s.d.]. Langues : Français ( fre) | Catégories : | e-Management et Business Intelligence
| | Index. dĂ©cimale : | 2119/19 | | RĂ©sumĂ© : | The human face plays an important role in our social interaction, conveying people’s
identity. Using the human face as a key to security, face recognition technology
has received significant attention in the past several years due to its potential for a
wide variety of real-life applications.
The goal of this project is to study this technology and build a robust Face Recognition
system that can be compared to Human-Level performance. To achieve this
goal, we first had to go into the background underlying this technology, by searching
for and analyzing state-of-the-art algorithms. Hence, a face recognition system is
always a combination of two main concepts, face detection and face recognition, as
a consequence this project is tackling both concepts.
A survey of available machine learning algorithms will be presented by going
through the analysis of recently published benchmarks. Based on these benchmarks
results, a list of potential algorithms is selected, as well as two evaluation data-sets.
This selection allows to structure our own experiments, by taking those algorithms
and testing them with different evaluation protocols, and determine which algorithms
perform better and find out relevant algorithms for our system, that is to
say MTCNN [1] for detection and FaceNet [2] for recognition.
Finally, to have more reliable results with regards to this project, A prototyping
use case has been built in shape of biometric solution for payment. This work
describes all the steps required to design the use case. Then, we report on the development
details of our system and also its deployment process into the prototype.
The characterization of hardware and the description of software is presented. |
|
211/01 
Faire une suggestion Affiner la recherche
Deployment of a Face Recognition System within a practical context / Abdelkarim AZOUZI
