norse.torch.functional.receptive_field

A module for creating receptive fields.

Functions

calculate_normalization(dx, scale[, gamma])	Calculates scale normalization for a spatial receptive field at a given directional derivative Lindeberg: Feature detection with automatic scale selection, eq.
covariance_matrix(sigma1, sigma2, phi)	Creates a 2-dimensional covariance matrix given two variances and an angle for the major axis.
derive_kernel(kernel, angle)	Takes the spatial derivative at a given angle
derive_spatial_receptive_field(field, scale, ...)	Derive spatial receptive field at a given angle and scale with respect to a list of derivatives.
derive_spatial_receptive_field_single(field, ...)	Calculate the derivative of a single spatial receptive field at a given angle and scale with respect to x and y derivatives.
gaussian_kernel(size, c, x, y[, domain])	Efficiently creates a differentiable 2d gaussian kernel.
spatial_parameters(scales, angles, ratios, ...)	Combines the parameters of scales, angles, ratios, xand y coordinates of the center of the rf and derivatives as cartesian products to produce a set of parameters for spatial receptive fields.
spatial_receptive_field(scale, angle, ratio, ...)	Creates a (size x size) receptive field kernel at a given scale, angle and ratio with respect to x and y derivatives.
spatial_receptive_fields_with_derivatives(...)	Creates a number of receptive fields based on the spatial parameters and size of the receptive field.
spatio_temporal_parameters(scales, angles, ...)	Combines the parameters of scales, angles, ratios and derivatives as cartesian products to produce a set of parameters for spatial receptive fields.
temporal_scale_distribution(n_scales[, ...])	Provides temporal scales according to [Lindeberg2016].