/3DMNIST

3D version of the MNIST database of handwritten digits

Primary LanguageJupyter Notebook

3DMNIST

A 3D version of the MNIST database of handwritten digits

3D MNIST

CONTENT

The dataset includes 5000 (train), 1000 (valid) and 1000 (test) 3D point clouds stored in HDF5 file format.

The files can be found in the data subfolder, containing 5000(train_small), 1000(valid_small) and 1000(test_small) HDF5 groups

Each group is named as its corresponding array index in the original mnist dataset (also included in the data subfolder) and it contains:

  • "points" dataset: x, y, z coordinates of each 3D point in the point cloud.
  • "normals" dataset: nx, ny, nz components of the unit normal associate to each point.
  • "img" dataset: the original mnist image.
  • "label" attribute: the original mnist label.

In the 3DMNIST notebook you can find the code used to generate the dataset.

You can use the code in the notebook to generate a bigger 3D dataset from the original.

CONTRIBUTIONS

In the contrib subfolder you can find examples on how to read/write the HDF5 files in Python and on how to generate global/local features to train predictive models.

New contributions are welcome via pull_request. Please try to follow the schema of existing contributions.

Current contributions:

  • voxelgrid: An example on how to generate a grid of voxels to extract a global feature vector and use that vector for training a linear model.