What Are Annotations?

Annotations are metadata labels that describe and categorize images or objects within images. They help structure visual datasets, making them searchable, analyzable, and useful for AI training.

Common annotation types:

  • Image Annotations: Assign class labels to entire images, helping categorize datasets.
  • Object Annotations: Label individual objects within images using bounding boxes, improving model accuracy.

Why Are Annotations Important?

Annotations play a critical role in machine learning and AI model training by:
✔ Improving data organization for easier retrieval and filtering
✔ Enhancing search-ability within datasets
✔ Supporting object detection and classification tasks
✔ Increasing model accuracy with high-quality labeled data


Supported Formats

  1. Parquet / CSV
  2. JSON (COCO format)

Your annotation file must be named exactly as one of the following:

  • annotations.json
  • image_annotations.csv
  • object_annotations.csv
  • image_annotations.parquet
  • object_annotations.parquet

Important: Annotations must be added when creating a dataset.


How to Import Annotations

  1. Upload your annotation file during dataset creation.
    Files can be uploaded from your local machine and S3 bucket.
  2. Ensure that your file follows the required format and has the correct name.


For more details about creating your annotation file, visit - Preparing annotation file