Reconstructed Slides: The Web Browser Is All You Need - Paul Klein IV, Browserbase

Source Video

The Web Browser Is All You Need - Paul Klein IV, Browserbase

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 160.06

• Source frame: slide-002.jpg
• Crop: full [0, 0, 960, 540] score 175.89

• Source frame: slide-003.jpg
• Crop: full [0, 0, 960, 540] score 176.63

• Source frame: slide-004.jpg
• Crop: full [0, 0, 960, 540] score 176.36

• Source frame: slide-005.jpg
• Crop: full [0, 0, 960, 540] score 175.96

• Source frame: slide-006.jpg
• Crop: full [0, 0, 960, 540] score 153.78

• Source frame: slide-007.jpg
• Crop: full [0, 0, 960, 540] score 166.35

• Source frame: slide-008.jpg
• Crop: full [0, 0, 960, 540] score 177.09

• Source frame: slide-009.jpg
• Crop: full [0, 0, 960, 540] score 167.22

• Source frame: slide-010.jpg
• Crop: full [0, 0, 960, 540] score 176.41

• Source frame: slide-011.jpg
• Crop: full [0, 0, 960, 540] score 159.39

Dense Scene-Detected Slide Candidates

• youtube YRGjll7uu5w dense slides