STIVL dataset was collected using an Orbbec Astra2 camera with a handhold manner. this dataset records RGB-D images and corresponding camera poses of four different indoor environments. We utilize the dense 3D reconstruction system ROSEFusion to get the pose label. Note that the RGB camera and the depth camera have been transferred to the same coordinate system in advance. The dataset is available at here.
For each scene, four sequences are recorded, in which three sequences are used for training and one sequence for testing.
- CORNER 1:a corridor with spatial extent
$2m^3$ , including 3000 images for training and 1000 images for test. - CORNER 2:a corridor with spatial extent
$2m^3$ , including 3000 images for training and 1000 images for test. - ELEVATOR LOBBY 1:a room with spatial extent
$3m^3$ , including 3000 images for training and 1000 images for test. - ELEVATOR LOBBY 2:a room with spatial extent
$9m^3$ , including 3000 images for training and 1000 images for test.