Authors :
Mukamana Florentine; Ngaruye Innocent; Vuganeza Patrice; Nyesheja Muhire Enan; Sangwa Octave; Maniriho Claudien
Volume/Issue :
Volume 9 - 2024, Issue 1 - January
Google Scholar :
http://tinyurl.com/3uhm2ddc
Scribd :
http://tinyurl.com/37msjvbt
DOI :
https://doi.org/10.5281/zenodo.10618701
Abstract :
Insurance companies continue to struggle to
deliver effective customer service and automated client
interactions in the modern digital world. This problem
becomes evident when serving customers who have lost
their physical documents or when fraud occurs. In
response to these challenges, a face recognition system
was created to enhance customer identification and
mitigate fraud risks within insurance companies. The
system matches customer faces with the information
stored in three databases, thereby enhancing customer
service and security measures. The Multi-Task
Cascaded Convolutional Neural Networks (MTCNN)
and VGGface models were used for face detection and
recognition, while SQLite databases were employed for
storing customer information. The streamlit Python
package was used to input the uploaded image or
camera image and the visualization of system output.
The developed face recognition system offers insurance
companies a means to identify customers through image
recognition to improve customer service, increase their
efficiency in their activities and avoid frauds that may
lead to these insurance companies’ bankruptcy. This
system can be seamlessly integrated with existing
systems and applied in multiple areas, including
customer service, security, and access control. Overall,
it provides a practical solution for insurance companies
to improve performance and organizational culture
while reducing the risk of fraud. The accuracy of the
developed recognition approach is 96% which is higher
when compared to other existing approaches.
Keywords :
Convolution Neural Network, Face Recognition, Machine Learning, Feature Extraction, Deep Learning, Face Matching.
Insurance companies continue to struggle to
deliver effective customer service and automated client
interactions in the modern digital world. This problem
becomes evident when serving customers who have lost
their physical documents or when fraud occurs. In
response to these challenges, a face recognition system
was created to enhance customer identification and
mitigate fraud risks within insurance companies. The
system matches customer faces with the information
stored in three databases, thereby enhancing customer
service and security measures. The Multi-Task
Cascaded Convolutional Neural Networks (MTCNN)
and VGGface models were used for face detection and
recognition, while SQLite databases were employed for
storing customer information. The streamlit Python
package was used to input the uploaded image or
camera image and the visualization of system output.
The developed face recognition system offers insurance
companies a means to identify customers through image
recognition to improve customer service, increase their
efficiency in their activities and avoid frauds that may
lead to these insurance companies’ bankruptcy. This
system can be seamlessly integrated with existing
systems and applied in multiple areas, including
customer service, security, and access control. Overall,
it provides a practical solution for insurance companies
to improve performance and organizational culture
while reducing the risk of fraud. The accuracy of the
developed recognition approach is 96% which is higher
when compared to other existing approaches.
Keywords :
Convolution Neural Network, Face Recognition, Machine Learning, Feature Extraction, Deep Learning, Face Matching.
