Academia.eduAcademia.edu

Figure 2 – uploaded by Stephane bres

See full PDFdownloadDownload figure
visual saliency of text blocks in the gray level arrow. From the right down corner to the left up corner there is a progression in the visual saliency of text entities. The bold and big titles are represented in the left up corner whereas the standard text paragraphs are found in the right one. The increase of complexity and visibility in the 2D-graph induces this phenomenon.

Figure 2 visual saliency of text blocks in the gray level arrow. From the right down corner to the left up corner there is a progression in the visual saliency of text entities. The bold and big titles are represented in the left up corner whereas the standard text paragraphs are found in the right one. The increase of complexity and visibility in the 2D-graph induces this phenomenon.

Related Figures (6)

Fig.2. Visual content analysis of a page extracted from the Oulu) MT database with the corresponding 2D-graph (complexity, visibility).
Fig.2. Visual content analysis of a page extracted from the Oulu) MT database with the corresponding 2D-graph (complexity, visibility).
Fig. 5. Automaton process for the page labeling. identified structural layout elements. In that way, we propose to evaluate the visual similarity between two pages by comparing their visual characteristics. This similarity metric is presented in the following section.
Fig. 5. Automaton process for the page labeling. identified structural layout elements. In that way, we propose to evaluate the visual similarity between two pages by comparing their visual characteristics. This similarity metric is presented in the following section.
where A,.. is the gray level of pixel at (r,c) position. C gives a value between 0 and 1.
where A,.. is the gray level of pixel at (r,c) position. C gives a value between 0 and 1.
We will use now the previous system to perform query by example and the automatic page classification. We must recall that these tasks are based on visual similarity and not on semantic content which 1s not analyzed at all. Fig.7 Results obtained for the query by example.
We will use now the previous system to perform query by example and the automatic page classification. We must recall that these tasks are based on visual similarity and not on semantic content which 1s not analyzed at all. Fig.7 Results obtained for the query by example.
Related topics:
Computer ScienceDocument ClassificationQuery by Example
580 California St., Suite 400
San Francisco, CA, 94104
© 2025 Academia. All rights reserved