ikomia on Pypi

Build and deploy Computer Vision solutions with a few lines of code

About The Project

At Ikomia, we deeply believe that sharing scientific knowledge is the key to success. This belief drives us to make research-based algorithms ready-to-use for developers through our Python library (Ikomia API) and desktop software (Ikomia STUDIO).

Our main goal is to convert existing Python code into easily accessible and deployable algorithms. This approach enables us to integrate repositories from researchers, labs, and renowned frameworks like OpenCV, Detectron2, OpenMMLab and Hugging Face. Within a unified framework, developers can craft workflows and seamlessly blend these cutting-edge algorithms, streamlining the integration process by eliminating its complexities.

By democratizing advanced technologies, we strive to accelerate innovation in AI, computer vision, and machine learning. Our platform empowers developers to leverage cutting-edge algorithms effortlessly, thereby fostering collaboration and transformative applications.

from ikomia.dataprocess.workflow import Workflow
from ikomia.utils.displayIO import display

# Init your workflow
wf = Workflow()

# Add YOLO and connect it to your input data
yolov7 = wf.add_task(name="infer_yolo_v7", auto_connect=True)

# Run directly on your image
wf.run_on(url="https://raw.githubusercontent.com/Ikomia-dev/notebooks/main/examples/img/img_fireman.jpg")

# YOLO output image with bounding boxes
display(yolov7.get_image_with_graphics())

And finally, you can also export your results as JSON files.

# Get all object detection
json_results = yolov7.get_results().to_json()
print(json_results)

Usage 2 : Pose Estimation Example

from ikomia.dataprocess.workflow import Workflow
from ikomia.utils.displayIO import display

# Init your workflow
wf = Workflow()

# Add YOLO and connect it to your input data
yolov7 = wf.add_task(name="infer_mmlab_pose_estimation", auto_connect=True)

# Run directly on your image
wf.run_on(url="https://raw.githubusercontent.com/Ikomia-dev/notebooks/main/examples/img/img_fireman.jpg")

# YOLO output image with bounding boxes
display(yolov7.get_image_with_graphics())

Usage with the `ik` auto-completion system

ik is an auto-completion system designed to help developers to find available algorithms in Ikomia HUB. See the documentation for more explanations here.

from ikomia.dataprocess.workflow import Workflow
from ikomia.utils import ik
from ikomia.utils.displayIO import display

wf = Workflow()

yolov7 = wf.add_task(ik.infer_yolo_v7_instance_segmentation(), auto_connect=True)

# wf.run_on(path="path/to/your/image.png")
wf.run_on(url="https://raw.githubusercontent.com/Ikomia-dev/notebooks/main/examples/img/img_dog.png")

display(yolov7.get_image_with_graphics())
display(yolov7.get_image_with_mask())
display(yolov7.get_image_with_mask_and_graphics())

Export your workflow

from ikomia.dataprocess.workflow import Workflow
from ikomia.utils import ik

wf = Workflow("Dog instance segmentation with YOLOv7")

yolov7 = wf.add_task(ik.infer_yolo_v7_instance_segmentation(), auto_connect=True)
filter = wf.add_task(ik.ik_instance_segmentation_filter(categories="dog", confidence="0.90"), auto_connect=True)

wf.save("path/to/your_workflow.json")

Once you export your workflow, you can share it with others or use it with Ikomia STUDIO.

Examples

You can find some notebooks here.

We provide some Google Colab tutorials:

Notebooks	Google Colab
How to make a simple workflow
How to run Neural Style Transfer
How to train and run YOLO v7 on your datasets
How to use Detectron2 Object Detection

@misc{DeBa2019Ikomia,
  author =       {Guillaume Demarcq and Ludovic Barusseau},
  title =        {Ikomia},
  howpublished = {\url{https://github.com/Ikomia-dev/IkomiaAPI}},
  year =         {2019}
}