Computer Vision

Label images in-house for image annotation tasks such as object detection, image segmentation and image classification. Use Prodigy's fully scriptable back-end to build powerful active learning workflows by putting your model in the loop.

 skateboard  person  person 
source: Unsplash by: Kirk Morales url:

Continuously update and evaluate your image models

If you're running a model for object detection, image segmentation or image classification in production, the examples your model is classifying will be changing constantly. Prodigy makes it easy to validate your model's predictions, and correct its mistakes.

Recognize new objects and categories

Prodigy now also ships with beta support for manual image labeling. Stream in your images from a directory or even a custom Python function, and mark objects or draw polygon shapes using an intuitive, browser-based interface. Your annotations can be exported as a simple JSON file with pixel coordinates, making it easy to integrate the data into the rest of your pipeline.

def custom_recipe(dataset, image_dir):
    stream = Images(image_dir)
    stream = fetch_images(stream)
    model = load_your_model()
    return {
        'dataset': dataset,
        'stream': prefer_uncertain(model(stream)),
        'update': model.update,
        'view_id': 'image'

Plug in your own models

Custom recipes let you integrate machine learning models using any framework of your choice, load in data from different sources, implement your own storage solution or add other hooks and features. No matter how complex your pipeline is – if you can call it from a Python function, you can use it in Prodigy.

Put a human in the loop

Even if your data isn't sensitive or safety-critical, a little bit of review can go a long way – especially when deploying a new model for the first time. Prodigy makes it easy to create custom workflows for data curation and verification. It's also a handy way to perform a lot of little jobs: sometimes you just need to pick the best images from a collection of hundreds. With Prodigy, this only takes a few minutes.

Photo of a sandwich with salmon, red onions and herbs
source: Unsplash by: Monstruo Estudio url: