This is an RGB-D dataset for instance-level object detection, built by Kanzhi Wu, Centre for Autonomous Systems from University of Technology Sydney. The dataset consists 2 sets: 1) dataset for object detection under occlusion (scene 1) and 2) dataset for object detection under ambiguity (scene 2).

scene 1


Objects in this dataset are:

  • airpump
  • belvita_honey_nut
  • nutbar
  • sultana_bran
  • yoghurt_topps

The training object images can be downloaded from link and the scene dataset can be downloaded link.

A snapshot of the objects is scene1_object

scene 2


Objects in this dataset are:

  • belvita_chocolate
  • belvita_honey_nut
  • belvita_milk_cereals
  • master_chef
  • nutbar
  • oreo

The training object images can be downloaded from link and the scene dataset can be downloaded link.

A snapshot of the objects is scene2_object


To be continue.