Prototipo de Reconocimiento facial con detección de vida para el registro de asistencia al laboratorio de Software de la Carrera de Ingeniería en Sistemas/Computación de la Universidad Nacional de Loja.
The objective of this degree work was to develop a prototype of face recognition with liveness detection for attendance record at the Software laboratory of the Systems/Computer Engineering Career of the National University of Loja, as an alternative for the attendance record which does not require...
Salvato in:
| Autore principale: | |
|---|---|
| Natura: | bachelorThesis |
| Lingua: | spa |
| Pubblicazione: |
2023
|
| Soggetti: | |
| Accesso online: | https://dspace.unl.edu.ec/jspui/handle/123456789/27009 |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
Aggiungi Tag | Riassunto: | The objective of this degree work was to develop a prototype of face recognition with liveness detection for attendance record at the Software laboratory of the Systems/Computer Engineering Career of the National University of Loja, as an alternative for the attendance record which does not require the manipulation of instruments of community use. In addition, it shortens registration times and facilitates the creation of reports on the use of the laboratory. It is guided according to the Design Thinking framework. The empathy phase allowed to know the current attendance record process. In the second phase, the revelations gathered in the empathy phase were defined and the challenges for the development of the proposal were raised. In the third phase, ideas were collected as a solution to the challenges. In the fourth phase, the prototypes of the ideas obtained were made. Finally, in the fifth phase, the developed prototype was tested. The system consists of a web module that allows the administration of attendance record processes and through which face recognition is put into operation. The face recognition module was developed using the Face-recognition pre-trained classifier, which uses convolutional neural networks, and is supported by a liveness detection model developed with three-dimensional convolutional neural networks. The prototype developed does not require the physical manipulation of instruments for community use and has an accuracy of 95% in the predictions it makes. Moreover, it reduces attendance record time from a minimum of 2 estimated minutes used in the traditional way to 9.2 seconds average per person with this tool, and facilitates the creation of reports on the use of the laboratory in a timely manner. Keywords: Face recognition, liveness detection, computer vision, attendance record. |
|---|