Flexible and reliable tool to train machines to understand what they ‘see’
Open source image labeling tool for artificial intelligence training
As AI and machine vision are becoming more widespread, the burden of training neural networks for each machine learning task remains. Training AI for vision applications requires labeling large datasets of images, so that AI has sufficient information and examples to learn from to understand the context of what it ‘sees’. Needless to say, this can be extremely time-consuming and error-prone.
Anno is an open source tool for image labeling, offering many benefits:
- Standalone tool that can be used offline
- Powerful and flexible, but easily configurable
- Tried and tested in many real-world projects (and is an important part of our product suite).
Features
- Creates datasets to train AI for vision applications from images stored locally or remotely (via REST interface)
- Supports a wide range of marker types:
- Point
- Oriented point
- Axis aligned rectangle
- Rectangle with rotation
- Polygon
- Polyline
- Allows you to customize categories, marker shapes and behavior, and create dependencies between marker properties
- The dataset is a simple JSON file
- Dataset API – a Python library to read and manipulate dataset information