How It Works:
The redaction workflow is divided into four main stages:
-
Convert: Prepares the video frame for AI processing.
-
Detect: Identifies faces or other target objects.
-
Track: Follows detected objects across consecutive frames.
-
Redact: Applies the selected redaction effect to sensitive regions.
Key Features:
-
AI-Based Detection: Automatically detects faces and other supported objects.
-
Real-Time Redaction: Obscures sensitive regions while preserving the rest of the scene.
-
Object Tracking: Maintains consistent redaction as objects move through the video.
-
Flexible AI Models: Supports integration of custom detection models for different use cases.
-
Hardware Acceleration: Leverages technologies such as CUDA and TensorRT on NVIDIA platforms.
-
GStreamer Integration: Add redaction directly into existing multimedia pipelines.
-
C++ Library: Integrate the functionality into custom applications.
Example Use Case:
A surveillance system running on NVIDIA Jetson processes live camera streams and needs to protect personal identities before recording or transmitting video.
RidgeRun’s Object / Face Redaction Library detects and tracks faces, then applies redaction in real time before the video reaches downstream applications.
Use Cases:
-
Video surveillance
-
Smart cities
-
Automotive and transportation
-
Robotics and autonomous systems
-
Drones and UAVs
-
Video analytics
-
Privacy-focused AI applications
Supported Platforms:
-
NVIDIA Jetson
-
x86-64 Linux systems with NVIDIA GPUs
-
Other supported Linux-based platforms
Supported Formats:
-
RGB
-
RGBA
-
YUV
-
GREY
Why RidgeRun?
RidgeRun combines expertise in GStreamer, embedded Linux, AI, computer vision, and NVIDIA acceleration to provide real-time privacy solutions for demanding video applications.
Learn more about AI Based Object Redaction on our Developer Wiki:
Monthly Subscription
Hire RidgeRun at a flat monthly rate for a maximum of 8 hours per day, 40 hours per week each month. Get fast interaction with engineers working on your project, weekly 45 minute calls to update you on our progress, and the complete engagement and expertise of our engineering team.
Work With UsFAQs
It is a C++ library that uses AI-based detection to locate sensitive objects or people in visual media and obscure the detected regions to help protect privacy and identity.
The library can obscure detected regions using a redaction technique such as blurring.
Yes. The current implementation provides an AI model for face detection and redaction.
Customer Reviews
“I must say thus far yourself and the entire RidgeRun team have been a pleasure to work with. I realize what we have been asking for is on the demanding end of the spectrum. I am genuinely appreciative for all of the assistance you’ve provided us so far and have shown a notable amount of good-will in this effort. It has not gone unnoticed.”
Customer industry:
AudioVisual Devices
Customer Reviews
“It's been an absolute pleasure working with you, and everyone there at RidgeRun continues to impress us here. I've worked with many contractors and companies over my career and you are by far the most proficient and knowledgeable.”
Customer industry:
Drones
Customer Reviews
"As a side note: just to let you know, I'm really happy to be working with RidgeRun in general. It's pretty clear that you tell it like you see it, and know what you do and don't know; it's really refreshing. That's super rare on these kinds of technical engagements when everyone seems to either want to claim expertise they don't have, or are so cautious they are don't actually say anything useful"
Customer industry:
Drones and Robotics
Customer Reviews
"RidgeRun has been so good that part of me doesn't even want to write this testimonial—because if more companies discover them, I might have to share the amazing engineers I get to work with. But credit where credit's due: their team is sharp, responsive, and consistently delivers beyond expectations. They make tough technical challenges feel easy, and that's rare."
Ayman
Director of Software