Add document preprocessing submodules to OCR configuration

- Introduced new `SubPipelines` section in `ocr_config.yaml` to include `DocPreprocessor` settings.
- Added submodules for document orientation classification and unwarping, enhancing the OCR pipeline's ability to process various document formats and improve accuracy.
This commit is contained in:
Yury Kossakovsky
2025-08-29 17:31:47 -06:00
parent 12e2cf8ae1
commit d51c9c8ff6

View File

@@ -14,3 +14,18 @@ SubModules:
model_dir: null
batch_size: 1
score_thresh: 0
SubPipelines:
DocPreprocessor:
pipeline_name: doc_preprocessor
use_doc_orientation_classify: true
use_doc_unwarping: true
SubModules:
DocOrientationClassify:
module_name: doc_text_orientation
model_name: PP-LCNet_x1_0_doc_ori
model_dir: null
DocUnwarping:
module_name: image_unwarping
model_name: UVDoc
model_dir: null