L = -1/N ∑ [y_i log(ŷ_i) + (1 - y_i) log(1 - ŷ_i)]Attention(Q, K, V) = softmax(QK^T / √d_k)Vf(x) = 1 / (1 + e^{-x})∇ × E = -∂B/∂tH(p, q) = -∑ p(x) log q(x)E[X] = ∫ x f(x) dxP(A|B) = P(B|A)P(A) / P(B)w^T x + b = 0θ_{t+1} = θ_t - η ∇_θ J(θ)
[HASANTAVISION]
v2.0.4 // ONLINE
{ }

MAKE YOUR APP SEE

Hover over a demo to see it live.

Back to Home
Computer Vision // 2025-11

Face Recognition System

Demo: Left Panel ←
Source Code

Developed a high-accuracy facial recognition pipeline with anti-spoofing capabilities. The system achieves 99.8% accuracy on the LFW dataset and operates in real-time (30fps) on edge devices. Key features include: - ArcFace-based feature extraction - RetinaFace for robust face detection - Custom 3D passive liveness detection module - Milvus vector database integration for fast retrieval

Role

Lead AI Engineer

Technologies

PyTorchONNXTensorRTFastAPIReact

Use Cases & Advantages

- Seamless Access Control for High-Security Facilities - Automated Time & Attendance Tracking - Personalized Customer Experience in Retail - KYC Verification and Digital Onboarding Our Stack's Advantage: Unmatched accuracy paired with optimized TensorRT deployment guarantees <10ms inference time without compromising privacy or security.