DUTOptions

Options & Configuration · AiDotNet.Video.Options

Configuration options for the DUT (Deep Unsupervised Trajectory) video stabilization model.

Inheritance: ModelOptions → DUTOptions

For Beginners

DUT options configure the Deep Unsupervised Trajectory video stabilizer.

How It Works

References:

  • Paper: “DUT: Learning Video Stabilization by Simply Watching Unstable Videos” (Xu et al., ICCV 2022)

DUT learns stabilization in an unsupervised manner by watching unstable videos, predicting per-pixel flow fields for warping without requiring paired stable/unstable data.

Constructors

Constructor Summary
DUTOptions() Initializes a new instance with default values.
DUTOptions(DUTOptions other) Initializes a new instance by copying from another instance.

Properties

Property Summary
OnnxModelOptions OnnxOptions { get; set; } ONNX runtime options for inference mode.
VideoModelVariant Variant { get; set; } Model variant controlling capacity and speed trade-off.
double DropoutRate { get; set; } Dropout rate for regularization.
double LearningRate { get; set; } Learning rate for training.
double TemporalLossWeight { get; set; } Weight for the temporal consistency loss during training.
int NumFeatures { get; set; } Number of base feature channels in the flow estimation network.
int NumPyramidLevels { get; set; } Number of pyramid levels for coarse-to-fine flow estimation.
int NumResBlocks { get; set; } Number of residual blocks per pyramid level.
int TemporalWindowSize { get; set; } Temporal window size for multi-frame stabilization.
string ModelPath { get; set; } Path to the ONNX model file for inference mode.