Real-time object detection and three-dimensional tracking for a government mixed-reality experience
Delivered in 42 days and demonstrated live at the State Legislative Assembly.
A mixed-reality "Selfie Point" desktop application for the Lucknow Vidhan Sabha: live camera feed, real-time object detection and positional tracking compositing visitors into mixed-reality content, delivered in 42 days.
A public mixed-reality installation for the Uttar Pradesh Legislative Assembly, operated as a desktop application in a live government environment.
Rol: Computer-vision and application engineering
Delivered in partnership with Live Pixel and Mizzle Codes.
- From start to government demonstration
- 42 daysFrom start to government demonstrationFounder track record
- Detection and compositing, on local hardware
- Real timeDetection and compositing, on local hardwareNo cloud round-trips
- Public installation at the State Legislative Assembly
- LivePublic installation at the State Legislative AssemblyLucknow Vidhan Sabha
El sistema
Qué se construyó realmente
- Live camera feed captured and processed in real time
- YOLO-based object detection on every frame
- Spatial and positional tracking of detected subjects
- Mixed-reality composition of live subjects with rendered content
- Desktop application packaging for on-site operation
Por qué fue difícil
Complejidad de ingeniería
- Real-time detection and compositing on local hardware, no cloud round-trips
- A fixed 42-day window ending in a live public demonstration
- Operating reliably in an uncontrolled, high-footfall environment
Experiencia de producto
Del sistema entregado



Tecnología
Stack verificado
Solo se listan las tecnologías confirmadas en el código fuente, la documentación del proyecto o el despliegue en vivo.
- Python
- OpenCV
- YOLO object detection
- Real-time image processing
- Three-dimensional tracking
Entrega
Alcance de la responsabilidad
- Detection and tracking pipeline
- Mixed-reality composition
- On-site delivery
Relacionado
Sistemas cercanos
Automatic Crime Detector95.1%
A Django web application that analyses webcam and uploaded video frame by frame with trained YOLO models, captures anomalies with timestamps and sends automated email alerts, with remote web access.

American Way Health$430M+
A production voice-AI operation for American Way Health, built around Bland AI. The engineering contribution here is the production layer that keeps the agent calling, compliant and measurable at 400K+ calls per month.
Construya un sistema con una complejidad similar
Reserve una consulta gratuita de 30 minutos o escríbanos — siempre hablará directamente con la persona que construye su producto.
Techon Pixel · info@techonpixel.com
