Skip to content

Changelog

0.2.0 March 10, 2024

  • Benchmark submission portal on EvalAI
  • Updated benchmark instructions
  • Changed submission format to Run-Length Encoding (RLE) for improved efficiency. Thanks to ScanNet++ for the inspiration and to @elisabettafedele for the implementation
  • Nerfstudio format helpers
  • Fixed issues #5 (thanks to @DennisRotondi) and #8, #12 (thanks to @sankleta). This involved changes to the following assets: 468140/47334514/hires_wide (added), 464758/44358731/hires_depth (added), 420693/420693_annotations.json (modified), 420673/420673_descriptions.json (modified) and 467283/467283_annotations.json (modified)

0.1.0 October 10, 2024

  • Initial release
  • Dataset includes 315 scenes (200 training scenes, 30 validation scenes, 85 test scenes)
  • Toolkit code
  • Documentation page