Houdini 22.0 Nodes Geometry nodes

ML Deform geometry node

Uses an ML model to deform the skin of a character.

On this page
Since 21.0

Overview

This node uses a trained ML model to deform the skin of a character. It uses the same inputs as the Bone Deform SOP, with a fourth input for residual blend shapes. Each residual blend shape defines a displacement point in the skin mesh. Internally, the ML model predicts weights used to combine the residual blend shapes. These combined blend shapes are then used to apply a correction to the rest skin geometry.

Note

ML Deform assumes the model was trained on a data set that was created using the ML Pose Serialize SOP and its default settings - Mode set to Subskeleton and Include World Rotation turned on. The ML Deform SOP’s Joint Group parameter also needs to match the joint group used in the ML Pose Serialize SOP when creating the training data set. These requirements are met if the ML Train Deformer recipe was used for training.

See the Machine Learning documentation for more information.

Parameters

Model File

The trained model in ONNX format that takes as input a serialized pose and outputs a tuple of weights for the residual blend poses.

Reload Model

Forces a reload of the .onnx file.

Execution Provider

Determines which ONNX execution provider to use for inference. By default, the node attempts to pick the best available provider and prefers to use GPU acceleration.

Automatic

Chooses the best provider for the current system. This option prioritizes CUDA if installed, DirectML/CoreML as a fallback depending on the platform, and CPU inference if no GPU provider is available.

CPU

Perform inference using the CPU.

CUDA

Perform inference using CUDA/cuDNN. CUDA and cuDNN must be installed using the packages provided by NVIDIA.

DirectML

Only available on Windows. Performs inference using the Windows Direct Machine Learning library.

CoreML

Only available on macOS. Performs inference using Apple’s Core ML library.

Joint Group

The joint group whose pose to serialize for inclusion into the data set on which Model File was trained.

Joint Limits

When turned on, joint limits are applied to a pose before it goes into the model. Specify the name of the dictionary attribute that contains the joint limits.

Inputs

Skin to Deform

The skin that deforms by a proxy geometry, which does a tissue simulation.

Capture Pose

The capture pose for the capture weights on the bone surfaces.

Animated Pose

A deformed pose. This is not used and is passed straight to the corresponding output.

Residual Blend Shapes

The packed primitives that contain the residual blend shapes. The first residual blend shape always adds a weight of 1. The remaining residual blend shapes are scaled by the weights predicted by the ML model before they are applied.

Outputs

Deformed Skin

The skin that deforms by a proxy geometry, which has gone through a tissue simulation.

See also

Geometry nodes