Run Smart Prediction

Run Smart Prediction

To run Smart Prediction on an image:

  1. When in Data, select one or multiple images.
  2. Click the Run Smart Prediction button on the floating bar.
  3. In the pop-up, choose a prediction model: Panoptic segmentation, Instance segmentation, or Semantic segmentation.
  4. Click Run.

🚧

To run Smart Prediction, the image height and width should be between 257 px and 2048 px.

Prediction models

For Image Projects:

  • Cars: Predicts vehicles.
  • Instance segmentation (trained on COCO): Predicts your instance’s class label, bounding box, and binary mask. Use this model to detect countable objects, such as cats.
  • OCR (English text): Predicts texts in English.
  • Object detection (trained on COCO): Predicts objects using bounding boxes.

For Image (Legacy) Projects:

  • Instance segmentation (trained on COCO): Predicts your instance’s class label, bounding box, and binary mask. Use this model to detect countable objects, such as cats.
  • Panoptic segmentation (trained on COCO): Assigns all the pixels to classes. Use this model to detect countable objects, such as cars, and uncountable objects, such as the sky.
  • Semantic segmentation (trained on Cityscapes): Assigns all the pixels to classes. Use this model to annotate street scenes.