| Autor: |
Ebrahim, Sameh Shawki, Kamal, Yasser M., Fakhr, Whaled |
| Předmět: |
|
| Zdroj: |
International Journal of Simulation: Systems, Science & Technology; Mar2017, Vol. 18 Issue 1, p1-7, 7p |
| Abstrakt: |
In several cases there are many documents or images that were shredded either by hand or by a machine. Moreover the images will be highly rich in color than a document. In shredding images by hand there are many changed points that required a proposed model which is able to match these pieces. On the other hand shredding by machine has only four changed points for each shredded piece (at corners). This paper introduces a method to assemble shredded pieces to form the desired image which is seriously vital in many applications such as multimedia and security applications. The proposed method gets the interest points which have a change in the angle and stay unchanged for a long distance, also it will use three features to get the matching between pieces to increase the matching accuracy. Three features are distance between interest points, angle of change at interest points, and the boundary color at the interest points. The proposed method starts by getting the outer contour of each shredded piece, then detects the interesting points, then it tries to match between different pieces. It is clear that the number of points on the outer contour is very large, so detecting the interest points will improve performance for getting the right value of matching between each two shredded pieces. Finally, a complete graph is constructed for all shredded pieces which form a matching graph whose edges represent matching success value and vertices represent shredded pieces. The method scored an overall accuracy of matching of 85% compared to previous work and improved the overall performance by 14% compared with previous related work. [ABSTRACT FROM AUTHOR] |
| Databáze: |
Complementary Index |
| Externí odkaz: |
|
