Houdini 20.0 Nodes Geometry nodes

Topo Flow Sample geometry node

An operator that uses computer vision to identify distinct feature points in a given image file.

On this page

Overview

This operator finds the most prominant features of a scalar image provided as a file parameter. The search area may be restricted by providing an optional additional mask image. When two images are provided, the dimensions of both images must be equivalent. To use, first specify the maximum number of desired features. Then specify a quality measure that characterizes the strength of the feature, and a minmum separation distance that restricts features from being too close together. The output of the operator is a set of sample points over the image. These sample points may be used with Topo Flow for texture alignment when fitting a mesh to scan data.

Topo Flow Sample toolbar

When in the Topo Flow Sample view state, a toolbar appears across the top of the viewport. The controls help you visualize the result of the feature detection.

Label Features

Controls how feature labels should be displayed.

Off

No label numbers are displayed.

On

Displays an index number as the label.

Parameters

Texture

The image file to analyze for features.

Mask

An optional image file. When provided the search for features is restricted to the region of non-zero values from the mask.

Maximum Feature Size

The target number of feature points to find.

Minimum Spacing

The minimum voxel distance to maintain between feature points.

Quality Tolerance

Tolerance characterizing the minimal accepted quality of feature distinction. Features with a quality measure less than the tolerance are rejected.

Blurring Window Radius

Size of the blurring window used to analyze a voxel neighborhood.

Gradient Window Radius

Size of the window gradient used to analyze a voxel neighborhood.

Method

The evaluation method used to score features.

Corner Detection

Uses the Harris detection algorithm to find corner points.

Minimum Eigen Value

Uses Eigen Values from a local covariance matrix to find corner points.

Corner Weight

Available for Corner Detection parameter. Weight value used in the corner detection method.

Note

  • The image mask must have the same dimensions as the texture image.

See also

Geometry nodes