Vision & Robotics
AI Solutions
From contactless interfaces and privacy-preserving vision to autonomous field robots — intelligent systems that operate where people and machines meet.
the Physical World
This domain spans the interface between people and machines: multilingual support that runs on real equipment knowledge, contactless touch for shared terminals, privacy-first vision on existing infrastructure, and walking robots built for terrain wheels can't cross.
Dual Camera System
Two cameras compute 3D finger coordinates in real time. Touch is registered mid-air without physical contact — ideal for election equipment and public kiosks where hygiene matters.
Stereo vision computes 3D finger coordinates in real time. No depth sensor required.
Contactless touch interface. Optimized for election terminals and public kiosks.
Contactless Touch Interface — Stereo Vision
Real-time Masking
Edge-computing based personal information protection for existing CCTV infrastructure. No camera replacement needed — add an edge device and all processing happens on-site, eliminating any risk of personal data leakage.
Live De-identification across Multiple Cameras
Real-time Location & Contact Tracing
No camera replacement needed. All processing happens on-site, eliminating data leakage risk.
Real-time blur and masking of personal video data. Original footage restored only for authorized users.
Walking Robot
A mobile walking robot (hexapod / octapod) for challenging environments such as smart farms and vessels. Modular structure supports robotic arm attachment
— a versatile autonomous field platform.
Thrust 1200N · Max speed 500mm/s · Stroke 250mm · Accuracy ±0.1mm
Torque 150Nm · Angular velocity 360°/s · Range ±90° · Backlash <0.2°
Leg span 500×400mm · CoG 100mm below center · Mass dist. 45/55%
LiDAR FOV 270° · Camera FOV 120° · Electromagnet 500N · 24V max 200W
Prismatic + Revolute Axis Design
Differently
We understand the mechanism first. Then we build AI on top of it.
A common AI architecture spanning elections, medicine, and robotics. Insights from one domain improve another.
Dual cameras, LiDAR, stereo vision — we overcome single-sensor limitations with complementary modalities.
We build finished products that work in the field immediately — not research papers or tech demos.
