Facial Recognition Time Attendance System
Facial recognition Time Attendance System
Routinely used by both private and government organizations around the world Facial recognition Time Attendance Machine is Automatic face recognition can be used for legitimate and beneficial purposes, such as enhancing security. However, their power and omnipresence increase the potential negative impact that unfair methods can have on society. Such as discrimination against ethnic minorities. A necessary, but not sufficient, condition for legitimately deploying facial recognition algorithms is that their accuracy is the same for all demographic groups.
Guided by this purpose, researchers from the Human Pose Recovery and Behavior Analysis Group of the Center for Computer Vision – University of Barcelona. A research group led by Sergio Escalera, organized a challenge within the European Congress of Computer Vision (ECCV) 2020. As part of the challenge, participants submitted algorithms that had to carry out facial recognition tasks with the presence of attributes that could be confounding. The published results evaluated the precision and bias in gender and skin color taking into account the presence of these attributes.
The challenge was a success, as “it attracted 151 participants, who made more than 1,800 submissions in total. Which exceeded our expectations in terms of the number of participants and submissions,” explained Julio CS Jacques Jr., Researcher with the Scene Understanding group. Artificial Intelligence Lab (SUNAI), of the Computer Science, Multimedia and Telecommunications Studies of the Open University of Catalonia (UOC).
Participants used an unbalance image data set, simulating a real-world scenario in which artificial intelligence (AI) -based models had to be trained and evaluated on unbalanced data (considerably more white men than dark-skinned women). In total, the team worked with 152,917 images of 6,139 identities.
Images were scored for two protected attributes: gender and skin color; and for five legitimate attributes: age group (0-34, 35-64, 65 or older), head position (front, others), image source (still image, video frame), with or without glasses, and size of the bounding box.
The results obtained were very promising. The winning solutions exceeded 0.999 for accuracy, and also scored very low on the propose bias metrics. Which you can consider a step towards developing fairer facial recognition methods. The analysis of the top ten teams shows higher false positive rates for dark-skinned women and for samples in which both individuals wear glasses.
In contrast, there were higher false negative rates for fair-skinned men and for samples in which both individuals were under 35 years of age. “This was no surprise, because the adopted data set was not balanced in terms of different demographic attributes. However, it shows that general precision is not enough when the goal is to build fair facial recognition methods, as well as that future work on this topic must take into account precision and mitigation of bias at the same time.
Much more than facial recognition:
Face recognition is one technology well known but not widely used by famous drag of high price and complicated implementation low levels of reliability, difficulty integrating with other solutions and environments.
The ignorance of its possibilities.
We constantly receive information about unique and almost futuristic facilities from companies and organizations, and even so, this technology is not fully implemented on a regular basis.
The ignorance of its configuration possibilities together with the Data Protection law is another circumstance that slows down users; but on the other hand, what better way, by quick and solvent, to empower companies in the following areas:
- Personal access control.
- Security areas.
- Face mask.
- Reserved area.
- Social distance.
- Management of common areas.
- Covid protection equipment.
- Human resources.
- Resources, services and maintenance.
- You hire.
- Round’s control.
- Meeting room management.
- Dash Board
- Segmentations, age and gender.
- Anonymous demographic analysis.
- Management of spaces.
- Zones VIP’s.
- Noble areas.
- Restricted areas.
- Data center.
Designed to perform with high performance:
Our solution Artificial Intelligence for facial recognition is designed to operate economically with high performance and fast processing. To instantly detect and match millions of faces in near real time, even in difficult conditions where faces are moving, at different angles, in low light or partially darkened.
The system offers 99.87% accuracy and offers results 3-5 times faster than other facial recognition algorithms. Capable of detecting a moving face through live video and recognizing itself in less than 100 milliseconds. It also has a feature to improve precision. Which allows grouping the different forms of appearance of a person, including face masks, adding them manually.
Currently, with the widespread use of face masks, the system is “learning” to improve its logic of occlusion. Allowing health professionals to walk through safe areas without removing their protective elements, but maintaining their ability to recognize threats to safety.
American technology that goes beyond face detection, determining the age, gender, and even the mood of the subject.
This solution can be deployed on a single computer to monitor multiple IP cameras or scale to thousands of cameras in a distributed architecture, hosted locally or in the cloud.
At Vida IP we intend to democratize access to this technology and we have developed a great plan that includes information, procedures, integrations, demos, protocols, training, product and price, so that you can offer this technology to all customers, their factories, their buildings, as a standard move towards democratized digitization.
Driving schools will install biometric accesses in classroom:
The Government establishes new control measures in driving schools. Specifically, they will have to install a biometric access control and Facial recognition Time Attendance Machine both in classrooms and in vehicles for all CAP courses for road transport. Systems that must obtain authorization from the General Directorate of Land Transport.
Faced with this new regulation, many of them are already immerse in adaptation processes of their facilities. And are betting on solutions such as those offered by B + SAFE that are based on facial and biometric fingerprint recognition. The first allows each student to identify, even with a mask. And to easily and safely record the entrance and exit of the students, as well as measuring their temperature. A set of biometric solutions for driving schools and training centers. For drivers of certain vehicles intended for road transport with the Professional Aptitude Certificate (CAP).
To obtain the corresponding authorization from the General Directorate of Land Transport, and among other requirements. These centers must have a biometric access control system in order to facilitate the control of attendance at the courses that are taught, both in the part of training given in the classroom as well as that given aboard a vehicle.
Two biometric control systems
solutions B + SAFE is based on facial recognition and also biometric fingerprint recognition. The first allows each student to be identifies, even with a mask, and to easily and safely record the entrance and exit of students, as well as measuring their temperature.
The system records the attendance of each student to the CAP courses: entry and exit, course code and center identification; create different support groups; allows the records to be easily export in Excel or PDF format; and keep records safely.
The second system is based on fingerprint biometrics and offers the same levels of security and similar functionalities as the facial recognition system.