Skip to content

Latest commit

 

History

History

Folders and files

NameName
Last commit message
Last commit date

parent directory

..
 
 
 
 
 
 
 
 

README.md

Data Preparation

NYU-V2

The NYU-Depth V2 data set is comprised of video sequences from a variety of indoor scenes as recorded by both the RGB and Depth cameras from the Microsoft Kinect.

Download

Images and annotations from the official website, category mapping, etc. from the open source project.

Conversion

Download needed files into a folder. Then convert the files to the desired format, please run:

python gen_nyu.py -i <input_meta_dir> -o <output_dir>

input_meta_dir is the path where the raw data is stored. output_dir is the path where to save the converted data.

Note: The converted depth values are in millimetres, and save as uint16[0,65535] png, 0~65535(mm). Depth pixels where the depth is missing are encoded with 0.

SID(Stanford Indoor Dataset)

The 2D-3D-S dataset provides a variety of mutually registered modalities from 2D, 2.5D and 3D domains, with instance-level semantic and geometric annotations. It covers over 6,000 m2 collected in 6 large-scale indoor areas that originate from 3 different buildings. It contains over 70,000 RGB images, along with the corresponding depths, surface normals, semantic annotations, global XYZ images (all in forms of both regular and 360° equirectangular images) as well as camera information. It also includes registered raw and semantically annotated 3D meshes and point clouds. The dataset enables development of joint and cross-modal learning models and potentially unsupervised approaches utilizing the regularities present in large-scale indoor spaces. For more information on the dataset, visit the tools repo, the project site or the dataset wiki.

Download

The link will first take you to a license agreement, and then to the data.

The full dataset is very large at 766G. Therefore, we have split the data by area to accomodate a la carte data selection. The dataset also comes in two flavors: with global_xyz images (766G) and without (110G). Only need to download noXYZ.

Conversion

python gen_sid.py -i <input_meta_dir> -o <output_dir> -c <processer_nums>

input_meta_dir is the path where the raw data is stored. output_dir is the path where to save the converted data. processer_nums is the number of processes. Large volume of data requiring multi-process processing.

Note: Depth images are stored as 16-bit PNGs and have a maximum depth of 128m and a sensitivity of 1/512m. Depth pixels where the depth is missing are encoded with 65535.

Acknowledgments

The code to generate HHA comes from SUN RGB-D meta data repository.