class*args, **kwargs)

Class defining KPFCNN.

A model for Semantic Segmentation.

__init__(name='KPFCNN', lbl_values=[0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19], num_classes=19, ignored_label_inds=[0], ckpt_path=None, batcher='ConcatBatcher', architecture=['simple', 'resnetb', 'resnetb_strided', 'resnetb', 'resnetb', 'resnetb_strided', 'resnetb', 'resnetb', 'resnetb_strided', 'resnetb', 'resnetb', 'resnetb_strided', 'resnetb', 'nearest_upsample', 'unary', 'nearest_upsample', 'unary', 'nearest_upsample', 'unary', 'nearest_upsample', 'unary'], in_radius=4.0, max_in_points=100000, batch_num=8, batch_limit=30000, val_batch_num=8, num_kernel_points=15, first_subsampling_dl=0.06, conv_radius=2.5, deform_radius=6.0, KP_extent=1.2, KP_influence='linear', aggregation_mode='sum', density_parameter=5.0, first_features_dim=128, in_features_dim=2, modulated=False, use_batch_norm=True, batch_norm_momentum=0.02, deform_fitting_mode='point2point', deform_fitting_power=1.0, repulse_extent=1.2, augment_scale_anisotropic=True, augment_symmetries=[True, False, False], augment_rotation='vertical', augment_scale_min=0.8, augment_scale_max=1.2, augment_noise=0.001, augment_color=0.8, in_points_dim=3, fixed_kernel_points='center', num_layers=5, l_relu=0.1, reduce_fc=False, **kwargs)

Initialize self. See help(type(self)) for accurate signature.

augment_input(stacked_points, batch_inds, is_test)
big_neighborhood_filter(neighbors, layer)

Filter neighborhoods with max number of neighbors.

Limit is set to keep XX% of the neighborhoods untouched. Limit is computed at initialization

call(flat_inputs, training=False)

Calls the model on new inputs.

In this case call just reapplies all ops in the graph to the new inputs (e.g. build a new computational graph from the provided inputs).

Note: This method should not be called directly. It is only meant to be overridden when subclassing tf.keras.Model. To call a model on an input, always use the __call__ method, i.e. model(inputs), which relies on the underlying call method.

  • inputs – A tensor or list of tensors.

  • training – Boolean or boolean scalar tensor, indicating whether to run the Network in training mode or inference mode.

  • mask – A mask or list of masks. A mask can be either a tensor or None (no mask).


A tensor if there is a single output, or a list of tensors if there are more than one outputs.

get_batch_gen(dataset, steps_per_epoch=None, batch_size=1)

Method computing the batch indices of all points, given the batch element sizes (stack lengths).

Example: From [3, 2, 5], it would return [0, 0, 0, 1, 1, 2, 2, 2, 2, 2]

get_loss(Loss, logits, inputs)

Runs the loss on outputs of the model.

  • outputs – logits

  • labels – labels




Returns an optimizer object for the model.


cfg_pipeline – A Config object with the configuration of the pipeline.


Returns a new optimizer object.


Function called right before running inference.


data – A data from the dataset.


This function is called after the inference.

This function can be implemented to apply post-processing on the network outputs.


results – The model outputs as returned by the call() function. Post-processing is applied on this object.


Returns True if the inference is complete and otherwise False. Returning False can be used to implement inference for large point clouds which require multiple passes.


This function prepares the inputs for the model.


The inputs to be consumed by the call() function of the model.

preprocess(data, attr)

Data preprocessing function.

This function is called before training to preprocess the data from a dataset.

  • data – A sample from the dataset.

  • attr – The corresponding attributes.


Returns the preprocessed data

segmentation_inputs(stacked_points, stacked_features, point_labels, stacks_lengths, batch_inds, object_labels=None)
transform(stacked_points, stacked_colors, point_labels, stacks_lengths, point_inds, cloud_inds, is_test=False)

[None, 3], [None, 3], [None], [None]