AI Briefing
KOSign in

Netflix Deploys Automated Multi-Ratio Video Resizing System for Mobile Clips

·2026.10.27 01:01

Key point

The system generates 9:16, 4:5, and 1:1 crops in a single pass using a three-stage pipeline of shot detection, focus region discovery, and temporal reframing.

1 / 5

Details

  • Netflix built an automated multi-ratio video cropping system to support its 1.2B social media followers and the new vertical Clips feed in its mobile app.
  • The pipeline processes video in three stages: shot detection via neural networks and Active Speaker Detection, focus region discovery using face detections and saliency maps, and temporal reframing with heuristics for smooth camera movement.
  • The system generates 9:16, 4:5, and 1:1 crops in a single pass without additional computational overhead.
  • Integrated into internal Clips tools, the system allows editors to review and adjust crops, with XML export for external editors like Adobe Premiere.
  • Internal tests show the algorithm outperforms Adobe Premiere's Auto Reframe on a curated evaluation set.
computer-visionnetflixvideo processingaspect-ratioclipsautomated-editing
Read originalShare on Threads

This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.

Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.

Related

Netflix TechNetflix Pensive AI Covers 100% of Workload Failures, with Non-Users Showing 321% More Manual Log DebuggingNetflix TechBeyond API Compatibility: Building a Tool to Measure the Blast Radius of Library ChangesHacker NewsKrea 3 Launches: Frontier Image Editing Model Claims 10x Speed of Nano Banana ProRedditNew LoRA for Qwen3-VL-8B Adds Bounding Box Coordinates to Document ExtractionTLDR AINetflix Achieves >99.95% GPU Node Availability with Automated RemediationStability AI Research4Director Controls Video World Models with Rigid 3D Geometry