Meta: Video Analysis Pipeline

[satra]

Umbrella issue for video-level analysis capabilities.

Sub-issues

• Task: Support pose estimation on video #249 Support pose estimation on video (currently image-only, users must loop frames)
• Task [video analysis] face analysis #192 Face analysis on video
• Pipeline [video analysis] active speaker detection #194 Active speaker detection pipeline

Scope

Extend video analysis beyond single-image operations. Add video-level APIs that handle frame iteration internally, including pose estimation across video frames (#249), face analysis (#192), and active speaker detection (#194).