{ "version": "0.2.0", "git_url": "https://github.com/qurator-spk/eynollah", "tools": { "ocrd-eynollah-segment": { "executable": "ocrd-eynollah-segment", "categories": ["Layout analysis"], "description": "Segment page into regions and lines and do reading order detection with eynollah", "input_file_grp": ["OCR-D-IMG", "OCR-D-SEG-PAGE", "OCR-D-GT-SEG-PAGE"], "output_file_grp": ["OCR-D-SEG-LINE"], "steps": ["layout/segmentation/region", "layout/segmentation/line"], "parameters": { "models": { "type": "string", "format": "file", "content-type": "text/directory", "cacheable": true, "description": "Path to directory containing models to be used (See https://qurator-data.de/eynollah)", "required": true }, "dpi": { "type": "number", "format": "float", "description": "pixel density in dots per inch (overrides any meta-data in the images); ignored if <= 0 (with fall-back 230)", "default": 0 }, "full_layout": { "type": "boolean", "default": true, "description": "Try to detect all element subtypes, including drop-caps and headings" }, "tables": { "type": "boolean", "default": false, "description": "Try to detect table regions" }, "curved_line": { "type": "boolean", "default": false, "description": "try to return contour of textlines instead of just rectangle bounding box. Needs more processing time" }, "allow_scaling": { "type": "boolean", "default": false, "description": "check the resolution against the number of detected columns and if needed, scale the image up or down during layout detection (heuristic to improve quality and performance)" }, "headers_off": { "type": "boolean", "default": false, "description": "ignore the special role of headings during reading order detection" } }, "resources": [ { "description": "models for eynollah (TensorFlow format)", "url": "https://qurator-data.de/eynollah/2021-04-25/SavedModel.tar.gz", "name": "default", "size": 1483106598, "type": "archive", "path_in_archive": "default" } ] } } }