Ready-to-use SOTA algorithms for vision

A large collection of open-source state-of-the-art Computer Vision algorithms, with tested and reusable code.

280 algorithms in all areas of computer vision

280 algorithms in all areas of computer vision

Edge detection, image histogram, image thresholding, image denoising, image classification, object detection, semantic, instance and panoptic segmentation, pose estimation, text detection, text recognition, etc.

7 supported frameworks

OpenCV, Kornia, OpenMMlab, YOLO, Scikit, Detectron2, HuggingFace.

7 supported frameworks and counting
Yolo logo

Ready to use and editable

We packaged the best open source algorithms to make them universally compatible and editable with the Ikomia API and the STUDIO.

The HUB is the source. Choose where you want the upper hand:


Prototype with Python API


Prototype with STUDIO desktop app

Icon reverse

From research
to prototype

We select the best state-of-the-art algorithms for all Computer Vision engineers and data scientists.

Icon code

Also for inference
and training

Test the latest state-of-the-art deep learning algorithms with ease.

Icon settings

Plug annotated data with the dataset loaders

Whether it’s annotated with Roboflow, CVAT, or another software, load custom data with dataset loaders.

Join the community and start prototyping


Don’t see your question? Email

Are there prerequisites for using the HUB?

The HUB is the source of algorithms customized to be used in the API or STUDIO. Therefore there are no prerequisites.

Are all algorithms available in Ikomia HUB open source?

Yes all Ikomia algorithms are open source, with each its licence chosen by the author.

What does it cost?

The Ikomia HUB, API, and STUDIO are open source and free, and will always be!

Do I need to download the library?

Not at all. Just choose the algorithms that suit your needs in the API or STUDIO and they will download at once.

Does Ikomia support NLP?

No it doesn’t, Ikomia is Computer Vision centric.

Is there documentation?

No need for documentation to use the HUB since it is the algorithm source for API and STUDIO.

Can I modify the algorithms?

Yes, all algorithms are open source so you can easily edit the code.

How can I publish my algorithm in the HUB for the community to use and improve?

This is a work in progress, and you should be able to do so in a matter of months.

Can I apply the HUB’s training algorithms to all annotated data?

Use data annotated with Coco, Pascal voc, Via, Yolo, Cwfid, Wgisd and Wildreceipt.

Can I use the HUB’s algorithms with my tools?

These state-of-the-art algorithms have been standardized and packaged for the Ikomia API and STUDIO. That’s what enables them to be easily tested and trained with reusable code patterns and chained in the same workflow, regardless of their original framework. To use your tools, pick the original algorithms in their repos.