IRIS Suite · Video Enhancement
IRIS-SR: AI-Based Video Upscaling
AI-powered super resolution for real-time video upscaling on affordable FPGA hardware.
<1 month
Time to Market
Our automated design processes get your system up and running fast — so you can focus on your product, not individual components.
48×
Cost Reduction
IRIS-SR fits on a single mid-range FPGA. Prior to Heronic, running a 1080p @ 30 FPS upscaler required more than 6 high-end FPGA devices.
Drop-in
Integration
No compilers, no large runtimes — just a pure hardware solution that drops straight into your existing video processing pipeline.
See It In Action
IRIS-SR Running Live
Watch IRIS-SR performing real-time super resolution on live video. The demo showcases 4K upscaling from a 1080p source with sub-frame latency on a single mid-range FPGA device — no GPU, no CPU inference pipeline.
How It Works
The Super Resolution Pipeline
Visualising the 3× upscaling process. The original low-resolution pixel grid is spatially mapped into a larger footprint, then bilinear interpolation fills the gaps to produce a sharp, high-resolution output.
Use Cases
Target Applications
-
Region of Interest Enhancement
Cropped regions of a video can be zoomed in, recovering the original resolution and quality. ROI systems operate instantaneously and can handle multiple views at once.
-
Low Resolution Upscaling
With the majority of production equipment still at FHD, achieving 4K broadcast requires real-time upscaling. Our hardware increases quality up to what customers demand.
-
Video Wall Processing
Driving a high-resolution video wall from a low-resolution source requires powerful SR algorithms to ensure picture quality looks acceptable across all panels.
Target Markets
Our IP adds value across a wide range of industries.
Product Brief
Download the IRIS-SR Brief
Fill in your details and we'll send the product brief directly to your inbox.