qurator-spk
/
sbb_pixelwise_segmentation
mirror of https://github.com/qurator-spk/sbb_pixelwise_segmentation.git
1
0
Fork 0
Code Issues Releases Wiki Activity
100 Commits
2 Branches
0 Tags
841 KiB
Commit Graph

100 Commits (4150675621fb201975aeb40bed1e7553d2b80d24)
 

Author SHA1 Message Date
vahidrezanezhad 4150675621 early_erosion is added
vahidrezanezhad cca4d17823 new augmentations for patchwise training
vahidrezanezhad 5f456cf508 fixing artificial class bug
vahidrezanezhad c502e67c14 adding foreground rgb to augmentation
vahidrezanezhad 4f0e3efa2b early dilation for textline artificial class
vahidrezanezhad 9904846776 using prepared binarized images in the case of augmentation
vahidrezanezhad f31219b1c9 scaling, channels shuffling, rgb background and red content added to no patch augmentation
vahidrezanezhad 95bbdf8040 updating augmentations
vahidrezanezhad 7be326d689 augmentation function for red textlines, rgb background and scaling for no patch case
vahidrezanezhad 85dd59f23e update
vahidrezanezhad f4bad09083 save only layout output. different from overlayed layout on image
Clemens Neudecker b6bdf942fd
add documentation from wiki as markdown file to the codebase
vahidrezanezhad 59e5892f25 erosion rate changed
vahidrezanezhad 5fbe941f53 inference updated
vahidrezanezhad 30894ddc75 erosion and dilation parameters are changed & separators are written in label images after artificial label
b-vr103 c340fbb721 increasing margin in the case of pixelwise inference
b-vr103 f2692cf8dd brightness augmentation modified
vahidrezanezhad 9521768774 adding degrading and brightness augmentation to no patches case training
vahidrezanezhad 55f3cb9a84 printspace_as_class_in_layout is integrated. Printspace can be defined as a class for layout segmentation
vahidrezanezhad 647a3f8cc4 resolving typo
vahidrezanezhad c0faecec2c update inference
vahidrezanezhad 033cf6734b update reading order machine based
vahidrezanezhad 9358657a0d update config
vahidrezanezhad 743f2e97d6 Transformer+CNN structure is added to vision transformer type
vahidrezanezhad f1fd74c7eb transformer patch size is dynamic now.
vahidrezanezhad 2aa216e388 binarization as a separate task of segmentation
vahidrezanezhad 41a0e15e79 updating train.py nontransformer backend
vahidrezanezhad 815e5a1d35 updating train.py
vahidrezanezhad dc356a5f42 just defined graphic region types can be extracted as label
vahidrezanezhad b1d971a200 just defined textregion types can be extracted as label
vahidrezanezhad 1c8873ffa3 just defined textregion types can be extracted as label
vahidrezanezhad e25a925169
Update README.md
vahidrezanezhad b9cbc0edb7 replacement in a list done correctly
vahidrezanezhad 821290c464 scaling and cropping of labels and org images
vahidrezanezhad 4640d9f2dc modifying xml parsing
vahidrezanezhad 785033536a min_area size of regions considered for reading order detection passed as an argument for inference
vahidrezanezhad f6abefb0a8 reading order detection on xml with layout + result will be written in an output directory with the same file name
vahidrezanezhad 2e7c69f2ac inference for reading order
vahidrezanezhad 356da4cc53 min area size of text region passes as an argument for machine based reading order
vahidrezanezhad 29ddd4d909 pass degrading scales for image enhancement as a json file
vahidrezanezhad 5aa6ee0010 adding rest_as_paragraph and rest_as_graphic to elements
vahidrezanezhad 4e4490d740 machine based reading order training is integrated
vahidrezanezhad bf1468391a machine based reading order training dataset generator is added
vahidrezanezhad f5746011f6 use case printspace is added
vahidrezanezhad 4b7f7da07c use cases like textline, word and glyph are added
vahidrezanezhad 947a0e06f8 missing text types are added
vahidrezanezhad d687f5328f dynamic layout decorated with artificial class on text elements boundry
vahidrezanezhad f7dda078d2 page2label with a dynamic layout
vahidrezanezhad 5f06a02441 update requirements
vahidrezanezhad 2623113b0c
page to label enable textline new concept