Skip to content

rmsandu/IAT-image-annotation-website

Repository files navigation

IAT-image-annotation-website

Browser-based image editor for annotating several different structures on an image. Output is a PNG/SVG binary 2D segmentation mask. The output can be used for image segmentation task, such as machine learning semantic segmentation algorithms. This website provides basic drawing, editing, zooming and deleting functionality with free-hand pen-like drawing motion.

The website has 2 main functionalities Load File and Export File. A 2D image can be loaded from disk using Load File --> Upload Image . Several structures (mole, hair, pore, vessel, wrinkle, redness) can be then marked on the image. The brush size is adaptable. The strokes can be deleted (Draw Stroke/Erase Stroke toggle button) or undoed(Undo Stroke).

The website uses JavaScript mainly and it has been inspired by Raphael-Sketchpad.

Screenshots of the IAT website

Website after opening index.html. image

After annotating some structures: example_readme_2

Saving

Saving of the annotations can be done in 3 different format, PNG, SVG and JSON. Each annotation structure is exported individually in a single file. Exception is the JSON file, for which only a file is generated exclusive of the number of structures present.

MASK (Re)-Loading

Previous annotations can be reloaded onto the picture for editing or verification purposes. To accomplish this, first load the image (png) , then load the corresponding .json file from Load File --> Load Layers option.

Layer editing

More layers can be added. The existing ones can also be re-named.

About

Browser-based image sketchpad for annotating several structures on an image. Output is a JSON/PNG/SVG binary 2D segmentation mask that can be re-loaded.

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published