Reconstructed Slides: Trends Across the AI Frontier — George Cameron, ArtificialAnalysis.ai

Source Video

Trends Across the AI Frontier — George Cameron, ArtificialAnalysis.ai

Method

This deck is reconstructed from the existing video frame captures by detecting likely slide regions with OpenCV, cropping/upscaling those regions, deduplicating similar crops, and OCRing the cropped slide images locally. It is a cleaner companion to the full-stage frame deck.

Reconstructed Slides

• Source frame: slide-001.jpg
• Crop: full [0, 0, 960, 540] score 160.13

• Source frame: slide-002.jpg
• Crop: full [0, 0, 960, 540] score 157.49

• Source frame: slide-003.jpg
• Crop: full [0, 0, 960, 540] score 177.33

• Source frame: slide-004.jpg
• Crop: full [0, 0, 960, 540] score 176.97

• Source frame: slide-005.jpg
• Crop: full [0, 0, 960, 540] score 173.19

• Source frame: slide-006.jpg
• Crop: full [0, 0, 960, 540] score 173.51

• Source frame: slide-007.jpg
• Crop: full [0, 0, 960, 540] score 177.55

• Source frame: slide-008.jpg
• Crop: full [0, 0, 960, 540] score 174.02

• Source frame: slide-009.jpg
• Crop: full [0, 0, 960, 540] score 177.06

• Source frame: slide-010.jpg
• Crop: full [0, 0, 960, 540] score 166.76

• Source frame: slide-011.jpg
• Crop: full [0, 0, 960, 540] score 167.26

Dense Scene-Detected Slide Candidates

• youtube sRpqPgKeXNk dense slides