Most modern companies maintain an active directory of users, some also have records of their photos, as such updating and validating these photos is a time consuming task, especially in large companies.
Our system makes this process easier for both ends, the company and their associates: building a mobile application automates the process of sending or hand delivering a photo; creating a back-end service allows the company to tackle the problem of manually validating photos. By conjugating both these modules we create a system capable of fully automating this process in a safe way. Companies can now rest assured that every photo taken by a fellow associate respects their standards, and maintains their real identity. As for the University of Aveiro, it has an information system, which saves all the faces from students, professors, researchers and administration staff. This personal information is used by all services in the university, which helps in many tasks, namely by creating a profile page in a web portal, available to the public.
Until very recently, the Human Resources at University of Aveiro were responsible for the whole process of updating a photo. This includes receiving a photo from a user, validating the new photo, verifying the identity of both photos and then updating the photo in the information system. At the moment a system as deployed in a testing environment were only some members of the academic community are allowed to use. This system is already incorporated with idp.ua.pt.
The scope of our project resides in making a more refined and complete version of the already deployed service. This project will allow every member at the University of Aveiro to update their photo without direct interaction of the Human Resources department.
The system we designed is capable of defining metrics about the new photo. These metrics are relevant in many use cases. To this day it is possible to evaluate the position of the head, according to pitch, yaw and roll; distinguish if someones eyes are open or not; computing the quality of the image in accordance to BRISQUE, determine if someone is or isn’t wearing glasses or sunglasses and analyse the candidate eye of sight. All these parameters have ranges associated with them. This allows this service to be useful in multiple case scenarios, as well as by many companies and institutions.
By including a mobile application capable of detecting liveness we turn this system into something more robust as well as less frustrating to the user, since we clearly state that a photo of a face must be present within the photography frame. On the other hand, nowadays, smartphones are a part of our life and items that we carry along with us all the time, which makes updating a photo a less intrusive process since companies or institutions do not have the need to gather associates so that everyone's photo is up to date.
You can explore in detail the documentation of the Project through the Documentation Repository. Or you can visit our website to learn more.
If you would like, you can also look into the code behind the FotoFaces algorithms and architecture.
If you would like, you can also look into the code behind the Mobile Application frontend and Live Detection.
- Filipe Gonçalves - 98083
- Pedro Lopes - 97827
- Goncalo Machado - 98359
- Vicente Costa - 98515
- João Borges - 98155
- Demonstration
