VeriTex: Edge-Based Facial Liveness Verification for Electronic-Know-Your-Customer (eKYC) Process using Patch Based Majority Voting

Preventing face spoofing in eKYC systems is critical, but current anti-spoofing methods are often computationally intensive, relying on dedicated cloud servers. Therefore, deploying an edge implementation would save businesses costs by running the system directly on edge devices. In this paper, we introduce VeriTex, an edge-based Facial Liveness Verification approach designed to enhance eKYC processes against common spoofing attacks (print paper, print photo paper, and LCD screens) by using patch-based texture as the input feature and employing majority voting for decision-making. This system runs locally on an android phone, reducing cloud computing costs and latency by bypassing internet delays. The system is validated using False Acceptance Rate (FAR), False Rejection Rate (FRR), Accuracy, Precision, Recall, and F1 Score, showcasing state-of-the-art results on both newly formulated custom datasets and existing ones obtaining 15.28% FAR and 11.28% FRR on OULU NPU Dataset, and 4.3 HTER on Replay Attack Dataset.