Our product
Fennixar Vision
A desktop application for training vision AI models on your own images, without writing code.
Capabilities
Object detection
Locate parts, products, people or vehicles in an image.
Defect detection
Find scratches, cracks, paint defects, stains and other irregularities.
Counting
Count items in a tray, a box or a video frame.
Video tracking
Follow an object through a clip and label every frame.
Key features
- Automatic labelling
- Type a word such as "screw" or "bottle", and it draws the boxes for you to review.
- Training on good samples only
- It learns from good products alone, then flags anything that differs and shows where on the image.
- Four tasks in one application
- Move between images and video, and between tasks, without changing tools.
How it works
- Step 1
Label
Add your images and label them, automatically or by hand.
- Step 2
Train
Press start. It learns from your labelled examples.
- Step 3
Test
Try it on images it has never seen and see what it finds.
- Step 4
Use
Export the model file and run it in your own software.
Specifications
- Works with
- Images and video
- Runs on
- A Windows PC with an NVIDIA graphics card
- Internet
- Works fully offline. Your images stay on your PC.
- Cameras
- Any camera you already have
- Your model
- A standard file you export and keep, with no per-machine fees.
- Help
- Use it yourself, or our team trains it with you.
See Fennixar Vision on your own images.
Tell us what it should detect, and we will arrange a free live demonstration.