COCO Annotator vs Pixano

COCO Annotator is a web-based tool for creating image annotations used in training data for localization and object detection. It supports bounding boxes, polygons, segmentation masks, keypoints, and points, along with disconnected shapes treated as one instance, multiple labels on an image segment, and custom metadata. Users can import existing COCO datasets and export annotations in COCO format. Assisted options include DEXTR, MaskRCNN, Magic Wand, semi-trained model annotation, and Google Images dataset generation. The software includes user authentication and a REST API, with a Swagger interface at localhost:5000/api. It is self-hosted on Linux or Windows, and Docker with docker-compose is required because Docker is the only supported installation method. Documentation covers development and production builds, centralized datasets, and external access. The server uses Flask, Eventlet, and Gunicorn, while long-running requests go to RabbitMQ workers. The project recommends HTTPS to encrypt browser-to-site communication and invites users to its Discord community. It is listed as free under an MIT-licensed plan.

Pixano is a free, open-source tool for exploring and annotating computer-vision datasets, including collections of text, images, and video. Its annotation tools cover boxes, polygons, pixel masks, keypoints, cuboids, classification, and tracking; video labels can be propagated across time. Smart features include semantic search using models such as CLIP and segmentation using models such as SAM. Dataset formats such as COCO can be imported and exported, and Pixano uses Lance storage for dataset navigation. Its reusable Web Components can be assembled into custom apps. Pixano documents REST and Python APIs, along with Pixano Inference, a Ray Serve-based inference server with a Python client and REST API. Installation options documented by the project are pip in a Python virtual environment or an official Docker image. The supported Python range is 3.10 or later and earlier than 3.14. Pixano is licensed under CeCILL-C and is under active development, so its API may change. Three-dimensional point-cloud support is described as planned.