Update README.md

This commit is contained in:
Kate Webbink 2019-10-31 13:40:28 -05:00 committed by GitHub
parent 5013b326e7
commit 3c14f5de7b
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 14 additions and 9 deletions

View File

@ -1,24 +1,29 @@
# Collections-OCR
A script to batch collections label-images through OCR
Dependencies - make sure to install these libraries first:
## Dependencies
Make sure to install these libraries first:
- `magick` - to read in image files
- `tesseract` - to do OCR magic
- `stringr` - to split the OCR'ed lines to columns
## How to run the script:
Notes:
- this takes ~2 seconds per label-image
- this currently uses Tesseract's English ("eng"), German ("deu") and Latin ("lat") libraries.
- This takes ~2 seconds per label-image.
- Be mindful how many images you add to your "images" directory.
- Batch similar images together to facilitate the follow-up cleaning.
- This currently uses Tesseract's English ("eng"), German ("deu"), and Latin ("lat") libraries.
- Output is likely needs some [or many] follow-up/clean-up steps.
To run the script:
1. Add a folder named "images" to this script's directory
2. Add the images (JPG & JPEG) you'd like to OCR to that directory
3. Run the script (`Rscript ocrMangle.R`)
Output - a CSV named "ocrText-[Date-time].csv", containing these columns:
- "image" = filename for each JPG and JPEG
- "line_count" = number of lines in each OCR transcription
- "Line1" - "Line[N]" = text for each line in the OCR transcription.
- the number of "Line" columns will match the maximum number of lines as needed.
## Output
A CSV named "ocrText-[Date-time].csv", containing these columns:
- **"image"** = filename for each JPG and JPEG
- **"line_count"** = number of lines in each OCR transcription
- **"Line1"** - "Line[N]" = text for each line in the OCR transcription.
- the number of **"Line"** columns will match the maximum number of lines as needed.