PDF E14 analysis using DNN and a bunch of image processing hacks

assigned to @hpsaturn

changed target branch from main to first_round_cnn

added 1 commit

• bce7f235 - refactor to make digit region extraction more reliable

Compare with previous version

added 24 commits

• bce7f235...59903a02 - 4 commits from branch first_round_cnn
• 3956a290 - update segunda vuelta solo pdf
• 3795da6e - Merge branch 'test_vision' into 'devel'
• ff8689de - updated candidates names
• 4197bae5 - Merge branch 'first_round' of...
• df63da9e - Merge branch 'second_round' into 'main'
• 60e05ce3 - down de realibility
• 6bd04e75 - added reliability like environment variable
• 41941aa5 - fixed some issues with download method to twitter bot
• cd2adea4 - Merge branch 'fix_dowload_ex_issue' into 'devel'
• 63b51a06 - testing cnn first round processor
• a6bc9623 - testing new cnn method for reliability
• f25529a7 - refactor divipola from url to have code for cnn detection
• 0964cb90 - Merge branch 'devel' of...
• ce28e1e8 - migrated size of url field
• ee2c7fe8 - incremented size for url orgin info for reprocessing divpola
• bb849085 - implementation of ocr via cnn (new version)
• 8e97218b - fixed dependency for OpenCV libraries for cnn version
• 28107605 - fixed some excepcion issues
• 1c657ba1 - fixed vscode issue (undo) and restore url issue
• 4203ae14 - Merge branch 'feature/ocr-dnn_integration' into 'feature/ocr-dnn'

Compare with previous version

approved this merge request

mentioned in commit 717b4f1c

merged