Chinese-OCR / README.md
longmaodata's picture
Update README.md
073cd16 verified
metadata
extra_gated_prompt: Please fill out the following information to verify your academic identity.
extra_gated_fields:
  Name: text
  Phone Number: text
  Edu Email: text
  School: text
  Major: text
  Education level: text
  Enrollment date: date_picker
  Identity:
    type: select
    options:
      - Student
      - Teacher
task_categories:
  - image-to-text
language:
  - zh

Join the group

πŸš€πŸš€πŸš€πŸš€https://t.me/+Y5kL2iHis9A0ZWI1

βœ… No need to apply for direct access to other datasets

βœ… Mutual communication within the industry

βœ… Get more information and consultation

βœ… Timely dataset update notifications

Chinese OCR

Scene Types: Natural, Reshot, Screenshots

Collection Environments: Magazines, Newspapers, Books, Signage, Receipts, Maps, PPTs, Menus, Product Packaging, Train Tickets, Banners, Bulletin Boards, Cards

Lighting Distribution: Normal Lighting Conditions

Collection Devices: Web Scraping, Smartphone Photography

Collection Distance: 20cm to 500cm

Shooting Angles: Scenes with the phone placed flat on a table (front, left front, right front), handheld scenes, scenes taken with the phone mounted on a tripod (with different tilting angles)

Portrait and Landscape Mode: Includes vertical video shooting and horizontal picture taking

Data Characteristics: Rich shooting scenes covering multiple angles and lighting conditions

Data Formats: .jpg, .png

Image Resolution: 1280x720, 1496x720, 1920x1080, 2560x1440, 2244x1080

Dataset Size: 5027 images

Folder

image/png

Description

image/png