Versions:
VoTT 2.2.0, published by Microsoft, is an open-source electron-based Visual Object Tagging Tool engineered to streamline the creation of high-quality training datasets for computer-vision models. Designed for data scientists, machine-learning engineers, and annotation teams, the application supports frame-by-frame labeling of individual images or entire video sequences, allowing users to draw bounding boxes, polygons, and masks while assigning custom tags that map directly to object classes. Its project-centric workspace imports assets from local disks, Azure Blob Storage, or Bing Image Search, then exports annotations in COCO, Pascal VOC, YOLO, and TensorFlow CSV formats so the labeled data can be fed immediately into popular training pipelines such as Azure Custom Vision, YOLOv5, or Detectron2. Collaborative features include user-defined tag schemas, hot-key shortcuts, and an active-learning plugin that can leverage an existing model to pre-label new frames, significantly reducing manual effort on large volumes of footage. Because VoTT stores annotation metadata in a portable JSON project file, multiple annotators can work in parallel and later merge results without conflict. The tool runs cross-platform on Windows, macOS, and Linux, requires no installation beyond unpacking the release archive, and can be configured for light or dark themes to suit extended labeling sessions. Security-conscious teams appreciate that all processing occurs locally, ensuring sensitive imagery never leaves the premises. VoTT 2.2.0 is available for free on get.nero.com, with downloads provided via trusted Windows package sources such as winget, always delivering the latest version and supporting batch installation of multiple applications.
Tags: