Experimental determination of chosen document elements parameters from raster graphics sources
Visual appearance of documents and their formal quality is considered to be as important as the content quality. Formal and typographical quality of documents can be evaluated by an automated system that processes raster images of documents. A document is described by a formal model that treats a pa...
Đã lưu trong:
| Tác giả chính: | , , |
|---|---|
| Định dạng: | Bài viết |
| Năm xuất bản: |
2018
|
| Chủ đề: | |
| Truy cập Trực tuyến: | http://lrc.quangbinhuni.edu.vn:8181/dspace/handle/DHQB_123456789/3638 |
| Tags: |
Thêm thẻ
Không có thẻ, Hãy là người đầu tiên gắn thẻ bản ghi này!
|
| Tóm tắt: | Visual appearance of documents and their formal quality is considered to be as important as the content quality. Formal and typographical quality of documents can be evaluated by an automated system that processes raster images of documents. A document is described by a formal model that treats a page as an object and also as a set of elements, whereas page elements include text and graphic object. All elements are described by their parameters depending on elements’ type. For future evaluation, mainly text objects are important. This paper describes the experimental determination of chosen document elements parameters from raster images. Techniques for image processing are used, where an image is represented as a matrix of dots and parameter values are extracted. Algorithms for parameter extraction from raster images were designed and were aimed mainly at typographical parameters like indentation, alignment, font size or spacing. Algorithms were tested on a set of 100 images of paragraphs or pages and provide very good results. Extracted parameters can be directly used for typographical quality evaluation. |
|---|