Home
A web-based image annotation tool used to create training data for machine learning. It provides many distinct features including the ability to label an image segment (or part of a segment), track object instances, labeling objects with disconnected visible parts and efficiently storing the annotation and instances information in the well-know COCO format. The annotation process is delivered through an initiative and customizable interface similar to PhotoShop.
Computer vision systems have become significantly more efficient at tasks such as object detection, tracking, classification, and segmentation. Many of those systems utilize supervised learning techniques and train on large datasets of annotated images. As these datasets grow, however, annotating the images becomes increasingly more challenging and the efficiency of the annotation tool used becomes more essential. Unfortunately, very few annotation tools address this need for efficiency and versatility.
COCO Annotator provides many 2D image annotation features that other annotations tool fall short, this includes:
- Directly export to COCO format
- Segmentation of objects
- Useful API endpoints to analyze data
- Import datasets already annotated in COCO format
- Annotated disconnected objects as a single instance
- Labeling image segments with any number of labels simultaneously
- Allow custom metadata for each instance or object
- Magic wand/select tool