-
-
Dashboard
-
Reverse Tracer & ANPR
-
Facial Embeddings
-
Face tracking
About Vissora Desktop
The Inspiration
Vissora started with a simple thought: What if we could make our roads safer using intelligent video surveillance?
Coming from Odisha, I wanted to build something that could create a real impact in my own state. Road safety, traffic monitoring, and public security are challenges that require more than conventional CCTV cameras. Cameras are already installed across cities and highways, but extracting meaningful information from thousands of video feeds remains a significant challenge.
My vision is to bring this technology to the government of my state and explore how Vissora could contribute to smarter traffic monitoring, faster incident investigation, and safer roads.
What We Built
Vissora Desktop is a unified AI-powered video intelligence platform designed to manage and analyze up to 1,000 camera feeds through a single dashboard.
It goes beyond conventional CCTV monitoring by bringing multiple computer vision capabilities into one system:
- Face Recognition & Tracking: Identify and track faces across video feeds.
- Vehicle Detection & Tracking: Monitor vehicles and their movement across cameras.
- ANPR: Automatically recognize vehicle registration plates.
- Color & Appearance Search: Search for vehicles and objects using visual attributes.
- Timeline Intelligence: Store and organize detection events to make historical investigation easier.
- Centralized Monitoring: Manage multiple cameras through one unified dashboard.
How We Built It
We developed Vissora Desktop using Kotlin and Jetpack Compose, integrating computer vision technologies such as OpenCV and YOLO-based object detection.
The system combines video processing, object detection, tracking, and visual analytics to transform raw camera footage into structured information. Our focus has been on building an architecture that can handle multiple camera streams while keeping monitoring and investigation accessible through a single desktop interface.
What We Learned
Building Vissora pushed us to explore real-time computer vision, object tracking, video streaming, model inference, and desktop application architecture. We learned that building an intelligent surveillance platform is not just about detecting objects; it is about connecting detections across time, organizing information, and making it useful for real-world decisions.
Challenges
One of our biggest challenges was designing a system capable of handling multiple video streams without compromising performance. Integrating different detection and tracking capabilities, managing real-time video processing, and organizing detection events into searchable timelines required careful architectural decisions.
What's Next?
Vissora is more than a desktop application to us. We envision it as a foundation for smarter public infrastructure.
Our long-term goal is to present this platform to the Odisha state government and explore its potential applications in road safety, intelligent traffic management, and public security. We believe existing camera infrastructure can become significantly more valuable when combined with AI-powered video intelligence.
Log in or sign up for Devpost to join the conversation.