lightning_pose.models

lightning_pose.models.base Module

Base class for backbone that acts as a feature extractor.

Functions

normalized_to_bbox(keypoints, bbox)

convert_bbox_coords(batch_dict, ...)

Transform keypoints from bbox coordinates to absolute frame coordinates.

get_context_from_sequence(img_seq, ...)

Classes

BaseFeatureExtractor([backbone, pretrained, ...])

Object that contains the base resnet feature extractor.

BaseSupervisedTracker([backbone, ...])

Base class for supervised trackers.

SemiSupervisedTrackerMixin()

Mixin class providing training step function for semi-supervised models.

lightning_pose.models.heatmap_tracker Module

Models that produce heatmaps of keypoints from images.

Functions

upsample(inputs)

Upsample batch of heatmaps using interpolation (no learned weights).

Classes

HeatmapTracker(num_keypoints[, num_targets, ...])

Base model that produces heatmaps of keypoints from images.

SemiSupervisedHeatmapTracker(num_keypoints)

Model produces heatmaps of keypoints from labeled/unlabeled images.

lightning_pose.models.heatmap_tracker_mhcrnn Module

Models that produce heatmaps of keypoints from images.

Classes

HeatmapTrackerMHCRNN(num_keypoints[, ...])

Multi-headed Convolutional RNN network that handles context frames.

SemiSupervisedHeatmapTrackerMHCRNN(num_keypoints)

Model produces heatmaps of keypoints from labeled/unlabeled images.

UpsamplingCRNN(num_filters_for_upsampling, ...)

Bidirectional Convolutional RNN network that handles heatmaps of context frames.

lightning_pose.models.regression_tracker Module

Models that produce (x, y) coordinates of keypoints from images.

Classes

RegressionTracker(num_keypoints[, ...])

Base model that produces (x, y) predictions of keypoints from images.

SemiSupervisedRegressionTracker(num_keypoints)

Model produces vectors of keypoints from labeled/unlabeled images.

Subpackages