Completed
OCR Architecture
Class Central Classrooms beta
YouTube videos curated by Class Central.
Classroom Contents
Efficient OCR for Building a Diverse Digital History
Automatically move to the next video in the Classroom when playback concludes
- 1 Introduction
- 2 Digital Texts
- 3 Mass Digitization
- 4 Poor Performance
- 5 Sequence to Sequence Architecture
- 6 Efficient OCR
- 7 Digitization Tools
- 8 Modern OCR
- 9 FOCR vs Seek to Seek
- 10 CRN Architecture
- 11 OCR Architecture
- 12 Word Recognition
- 13 Models
- 14 Object Detection
- 15 Hard Negative Mining
- 16 Data Augmentation
- 17 OCR Benchmarks
- 18 Document Collections
- 19 Zero Shot Performance
- 20 Character Air Rate
- 21 Comparisons
- 22 Baseline Results
- 23 Japanese Results
- 24 Open Source OCR
- 25 ZeroShot Performance
- 26 Sample Efficiency
- 27 Oblations
- 28 Different Encoders
- 29 OCR at Scale
- 30 Custom Layout Model
- 31 Nonword Rate
- 32 Applications
- 33 Overall Data
- 34 Knowledge Graph
- 35 Supply Chain Network
- 36 Community Engagement
- 37 Training and Deploy
- 38 OCR encourages community engagement
- 39 Characters and words
- 40 Language extensibility
- 41 omitting the language model
- 42 decouple localization and recognition
- 43 limitations
- 44 extensions
- 45 fun example
- 46 conclusion