Returns the number of images for which the label appears.

Returns a 3-d numpy tensor with segmentation data for the ith image, with multiple layers represnting multiple lables for each pixel.

Creates a set of receptive field masks with uniform thresholds over a range of inputs.

Default implemtnation for segmentation_data, utilizing all_names.

Returns an English name for the jth label. If a category is specified, returns the name for the category-specific nubmer j.

Constructs a ResNet-101 model. Args: pretrained (bool): If True, returns a model pre-trained on ImageNet

Constructs a ResNet-18 model. Args: pretrained (bool): If True, returns a model pre-trained on ImageNet

Constructs a ResNet-34 model. Args: pretrained (bool): If True, returns a model pre-trained on ImageNet

Constructs a ResNet-50 model. Args: pretrained (bool): If True, returns a model pre-trained on ImageNet

Like numpy.zoom, but does not crash when the first dimension of the array is of size 1, as happens often with segmentations

Default implemtnation for segmentation_data, utilizing metadata and resolve_segmentation.

Returns a 2-d numpy matrix with segmentation data for the ith image, restricted to the given category. By default, maps all label nu

Applies a bicubic upsampling.

Upsampling utility functions

Applies a bilinear upsampling.