picdot

picdot is a software project focused on image annotation and data labeling that emphasizes a dot-based approach to marking features in images. It is commonly released as an open-source tool intended for computer vision workflows, where annotators place individual dots to indicate points of interest, with optional attributes attached to each dot.

The tool offers a lightweight editor, keyboard shortcuts, and a simple data model that supports per-dot labels, confidence scores, and grouping of dots into higher-level annotations. Annotated results can be exported in common interchange formats suitable for machine learning pipelines, and picdot provides basic integration hooks for exporting to standard schemas.

Designed to be cross-platform, picdot can run as a web application or as a desktop client, with optional server components to support project management and collaboration. Data can be stored locally or synchronized with a cloud backend, depending on configuration. The project emphasizes ease of use and rapid labeling for small to mid-sized datasets, where dot-level precision is preferred over polygonal outlines.

picdot has been used by researchers, educators, and hobbyists as a lightweight alternative to more complex annotation tools. Its development is community-driven, with ongoing updates to improve accessibility, extensibility, and interoperability. The project maintains documentation and a public repository that welcomes contributions, plug-ins, and workflow integrations.