TemporalAggregationType

Enums · AiDotNet.Enums

Specifies how frame-level features are aggregated into a single video-level representation.

For Beginners

When processing a video, each frame produces its own features. This enum controls how those per-frame features are combined into one representation for the whole video.

Fields

Field Summary
TemporalAggregationType LastFrame Takes only the last frame’s features as the video representation.
TemporalAggregationType MeanPooling Simple average (mean pooling) across all frame features.
TemporalAggregationType TemporalTransformer Uses a temporal transformer to learn attention-weighted aggregation across frames.
int value__