Giving feedback about computer vision performance (feedback loop)

If you encounter missed or wrongly classified road surface distresses, you can help improve the RoadAI computer vision by adding those images to our computer vision training data sets to increase defect detection accuracy.

  1. Make sure that the defect or object of interest is prominently displayed within the video frame.
    Use -1 and +1 to view one frame at a time.
  2. In the toolbar, select Tools.
  3. Under Annotations, switch on the Feedback loop.
  4. In the Select tags... field, type roadai_feedback and select Create tag "roadai_feedback".
  5. Optionally, use the field below to provide additional details for further clarification or context.
  6. Select Create.
    The tag roadai_feedback is associated with a data set specifically designed for image training. Any video frame with this tag will be incorporated into our collection of training images for computer vision algorithms to improve the accuracy and efficiency of our image recognition capability.