Sistema inteligente con Visión Artificial para el control de asistencia laboral en la Corte Superior de Justicia de Cañete, 2025
Loading...
Files
Date
2026-04-14
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Universidad Nacional de Cañete
Abstract
Esta investigación nace con la finalidad de implementar mejoras en el control de asistencia laboral de los trabajadores de la Corte Superior de Justicia de Cañete mediante el desarrollo de un sistema inteligente basado en visión artificial. Esta propuesta surge debido a que el registro tradicional demandaba que los trabajadores se desplazaran hasta sus oficinas para marcar su ingreso, lo que generaba pérdidas de tiempo, acumulación de personal en horas pico y, en algunos casos, accidentes por la prisa con la que intentaban llegar a su estación de trabajo. También se identificó que el sistema anterior no contaba con un monitoreo en tiempo real, lo que dejaba abierta la posibilidad de suplantaciones de identidad. Para el avance del sistema se aplicó la metodología ágil SCRUM y se trabajó con una muestra probabilística de 84 trabajadores administrativos seleccionados de una población total de 107. Durante la fase de evaluación se compararon los tiempos empleados en detección, reconocimiento y marcación antes y después de implementar el sistema, permitiendo observar los cambios producidos en cada dimensión. En detección, la media del pretest fue de 2.577 segundos y en el postest ascendió a 7.402 segundos, lo que representa un incremento del 187%. En reconocimiento se pasó de 7.177 segundos a 0.467 segundos, mostrando una disminución del 93.5% y reflejando un proceso mucho más rápido. En la marcación se observó el cambio más notable, ya que el tiempo promedio disminuyó de 331.624 segundos a 0.176 segundos, alcanzando una mejora del 99.95%. Con estos resultados se confirma que el sistema inteligente con visión artificial permitió agilizar el control de asistencia laboral y fortalecer la seguridad del proceso, ya que elimina la dependencia de registros manuales y ofrece una identificación más precisa y confiable.
The objective of this research was to improve attendance monitoring for employees of the Superior Court of Justice of Cañete by developing an intelligent system based on artificial vision. This proposal arose because the traditional system required workers to go to their offices to clock in, which led to wasted time, crowding during peak hours, and, in some cases, accidents due to the rush to get to their workstations. It was also found that the previous system did not have real-time monitoring, which left open the possibility of identity theft. The agile SCRUM methodology was applied to develop the system, working with a probabilistic sample of 84 administrative workers selected from a total population of 107. During the evaluation phase, the time spent on detection, recognition, and marking before and after implementing the system was compared, allowing changes in each dimension to be observed. In detection, the pretest average was 2,577 seconds and in the posttest it rose to 7,402 seconds, representing an increase of 187%. In recognition, it went from 7,177 seconds to 0.467 seconds, showing a decrease of 93.5% and reflecting a much faster process. The most notable change was observed in dialing, as the average time decreased from 331.624 seconds to 0.176 seconds, achieving an improvement of 99.95%. These results confirm that the intelligent system with artificial vision made it possible to streamline work attendance control and strengthen the security of the process, as it eliminates the dependence on manual records and offers more accurate and reliable identification.
The objective of this research was to improve attendance monitoring for employees of the Superior Court of Justice of Cañete by developing an intelligent system based on artificial vision. This proposal arose because the traditional system required workers to go to their offices to clock in, which led to wasted time, crowding during peak hours, and, in some cases, accidents due to the rush to get to their workstations. It was also found that the previous system did not have real-time monitoring, which left open the possibility of identity theft. The agile SCRUM methodology was applied to develop the system, working with a probabilistic sample of 84 administrative workers selected from a total population of 107. During the evaluation phase, the time spent on detection, recognition, and marking before and after implementing the system was compared, allowing changes in each dimension to be observed. In detection, the pretest average was 2,577 seconds and in the posttest it rose to 7,402 seconds, representing an increase of 187%. In recognition, it went from 7,177 seconds to 0.467 seconds, showing a decrease of 93.5% and reflecting a much faster process. The most notable change was observed in dialing, as the average time decreased from 331.624 seconds to 0.176 seconds, achieving an improvement of 99.95%. These results confirm that the intelligent system with artificial vision made it possible to streamline work attendance control and strengthen the security of the process, as it eliminates the dependence on manual records and offers more accurate and reliable identification.
Description
Keywords
sistema inteligente, visión artificial, reconocimiento facial, control de asistencia laboral, automatización.
Citation
Saavedra Medina Jose Luis , "Sistema inteligente con Visión Artificial para el control de asistencia laboral en la Corte Superior de Justicia de Cañete, 2025", Tesis para optar el grado de Ingeniero de Sistemas, Escuela de pregrado Facultad de Ingenieria de Sistemas. Universidad Nacional de Cañete. Cañete – Perú