TY - GEN
T1 - Computer Vision and Deep Learning-Based Model for Detecting Spoofed Faces in Images
AU - Salian, Gayathri P.
AU - Rao, Manasa K.
AU - Rashmi, M.
N1 - Publisher Copyright:
© The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.
PY - 2025
Y1 - 2025
N2 - In the modern era of smart applications, video data is critically important in various contexts. In most of these applications, cameras are frequently incorporated to facilitate authentication. As a result, face recognition is the biometric method most frequently employed to authenticate users in these applications. The vulnerability of face recognition systems to spoofing attacks grows in tandem with their increased usage. As a result, robust countermeasures are required. This paper presents an approach to face anti-spoofing through transfer learning and YOLOv8 optimization. Additionally, a custom dataset was constructed using images obtained from web cameras and an existing dataset to assess the proposed work’s real-time effectiveness. The proposed approach also adds a blurriness threshold during image capture to improve performance. With a mean Average Precision (mAP50) of 0.975, the experimental outcomes highlight the model’s effectiveness in detecting face spoofing.
AB - In the modern era of smart applications, video data is critically important in various contexts. In most of these applications, cameras are frequently incorporated to facilitate authentication. As a result, face recognition is the biometric method most frequently employed to authenticate users in these applications. The vulnerability of face recognition systems to spoofing attacks grows in tandem with their increased usage. As a result, robust countermeasures are required. This paper presents an approach to face anti-spoofing through transfer learning and YOLOv8 optimization. Additionally, a custom dataset was constructed using images obtained from web cameras and an existing dataset to assess the proposed work’s real-time effectiveness. The proposed approach also adds a blurriness threshold during image capture to improve performance. With a mean Average Precision (mAP50) of 0.975, the experimental outcomes highlight the model’s effectiveness in detecting face spoofing.
UR - https://www.scopus.com/pages/publications/85219188325
UR - https://www.scopus.com/pages/publications/85219188325#tab=citedBy
U2 - 10.1007/978-981-97-8336-6_2
DO - 10.1007/978-981-97-8336-6_2
M3 - Conference contribution
AN - SCOPUS:85219188325
SN - 9789819783359
T3 - Lecture Notes in Networks and Systems
SP - 15
EP - 26
BT - Data Science and Network Engineering - Proceedings ICDSNE 2024
A2 - Namasudra, Suyel
A2 - Kar, Nirmalya
A2 - Patra, Sarat Kumar
A2 - Taniar, David
PB - Springer Science and Business Media Deutschland GmbH
T2 - 2nd International Conference on Data Science and Network Engineering, ICDSNE 2024
Y2 - 12 July 2024 through 13 July 2024
ER -