Vision AI Laboratory focuses on computer vision, imaging AI, industrial inspection, and document intelligence.
Multimodal event and data correlation for semantic anomaly detection, risk scoring, situation classification, and LLM/VLM-based explanations and summaries.
High-precision object detection and multi-object tracking with event abstraction, behavior and crowd analysis, multi-sensor fusion, and lightweight streaming AI.
Efficient keyframe extraction algorithms that compress video content for optimized VLM input processing.
Time-series and simulation-based forecasting with high-speed AI search, Agentic RAG knowledge exploration, scenario-based decision logic, and NL operation.
Deep learning super-resolution targeting small or distant objects in low-resolution surveillance and industrial images.
Real-time video stabilization for drone footage, enabling stable visual analysis and motion compensation in aerial systems.
High dynamic range and adaptive contrast enhancement for low-light, IR, and thermal camera imaging systems.
AI-based image dehazing to restore object visibility under fog, haze, smoke, or rain conditions.
Automated X-ray defect inspection for semiconductor BGA packages and large-format cylindrical EV battery cells.
Low-dose X-ray denoising and high-speed 2D/3D reconstruction for minimizing radiation in industrial inspection.
AI detection of foreign particles in EV battery electrodes and food products using multi-modal sensing.
Fast 3D X-ray CT reconstruction and automated analysis for high-speed industrial defect inspection.
Intelligent document pipeline combining layout understanding and OCR for automated data extraction.
LLM and OCR based PII detection and de-identification for complex mixed-content document types.
Automated multi-level security classification for complex documents with mixed text, tables, and context.
OCR and semantic understanding of banners and outdoor advertisements for content classification; Public/Political/Commercial.
