GUI program that facilitates annotation of image data
- install all libraries from requirements.txt
pip3 install -r requirements.txt
- to launch the program, it needs to run the following command in a console:
python *\src\main.py
after that, all program's logs will appear in your console:
files removed: 0
labels removed 42
udpate: C:\projects\OCR\develop\to_valid\pe10.png
next: ('C:\\projects\\OCR\\develop\\to_valid\\pe1000.png',
udpate: C:\projects\OCR\develop\to_valid\pe1000.png наз-ся
...
fields:
img dir : path to the directory with images
labels : path to the *.tsv file with annotations
buttons:
save : save current changes
upload : upload images and labels into the program
remove : remove a image from the directory and the correspondent label in *.csv file
help : link to this page
next : the next pair (image, label)
previous : the previous pair (image, label)
hot keys:
shift + left arrow : previous
shift + right arrow : next
shift + space : remove
if spelling is not correct (i.e. word does not exist) it will make a signal.
Prepare a directory with images. For, example,
Prepare a *.tvs document with the list of image files. For, example
pe10.png
pe1000.png наз-ся
pe10001.png то есть
pe10003.png смещение
pe10005.png движения
pe10008.png Ищем
pe1001.png равномерно
pe10015.png
pe1002.png наиболее
pe10024.png больших
pe10026.png чему ещё
the second column is labels, it can be empty
Insert paths into correspondent fields and click "upload"
v. 0.3 (07.01.2022)
-- added an option to show only labels with spelling errors
v. 0.2 (01.01.2022)
-- spell checking for Russian added;
-- help link added;
-- GUI optimized.
v. 0.1 (23.12.2021)
uploaded
Error: UnicodeDecodeError: 'utf-8' codec can't decode byte 0xed in position 22: invalid continuation byte
Solution: open file with labels, copy its content and paste in the new one, save it; i.e. recreate a file with labels.