Antsy2d

in-browser point cloud annotation tool for instance-level segmentation using 2d projection. Original written by kyamagu with lightweight 3d point cloud adaptations by Alvin Wan.

• Label image regions with mouse.
• Written in vanilla Javascript, with require.js dependency (packaged).
• Pure client-side implementation of image segmentation.
• Fork introduces ability to label point cloud using 2d projection
• Includes support for KITTI dataset

A browser must support HTML canvas to use this tool.

Importing Data

You can use the automated script if you are using the KITTI dataset.

python to_antsy.py --kitti=path/to/KITTI

Otherwise, prepare a JSON file that looks like the following. The required fields are labels and imageURLs. The annotationURLs are for existing data and can be omitted. Place the JSON file inside the data/ directory.

{
  "labels": [
    "not drivable",
    "drivable"
  ],
  "imageURLs": [
    "data/images/test.png"
  ],
  "annotationURLs": [
    "data/annotations/test.png"
  ],
  "projectedURLs": [
    "data/projected/test.npy"
  ]
}

Then edit main.js to point to this JSON file. Open a Web browser and visit index.html. Once you're done annotating, click "save" to export. Drag the image into data/annotations. Once you're done, convert this data into labelled point clouds:

python from_antsy.py

Citation

The original author asks that future users cite the following:

@article{tangseng2017looking,
Author        = {Pongsate Tangseng and Zhipeng Wu and Kota Yamaguchi},
Title         = {Looking at Outfit to Parse Clothing},
Eprint        = {1703.01386v1},
ArchivePrefix = {arXiv},
PrimaryClass  = {cs.CV},
Year          = {2017},
Month         = {Mar},
Url           = {http://arxiv.org/abs/1703.01386v1}
}