Auto-reconstruction of Shredded Document based on Matching Models
Mengmeng Yu, Pinjie Ye, Shuoping Wang, Honghao Gao · International Journal of u- and e- Service Science and Technology · 2015
Shredding auto-reconstruction is a hot research topic in pattern recognition. The research progress can produce certain effect to various fields. The purpose of this paper is to study shredding auto-reconstruction based on regular shredded document from shredders, to obtain a practical and efficient splicing algorithm to auto-reconstruction of strip shaped shredded text documents and block shapedshredded text documents. For strips, this paper uses the pretreatment, the similarity matching model, combined with the optimalHamilton path algorithm, for which we get a good result with 100% correct rate and no human intervention. For blocks, first, this paper pretreats the fragments. And then uses the row cluster model to divide all debris to some rows, and then uses the similarity model with direct reverse matching model to achieve the shredding auto-restore in different rows. At last, we use line spacing matching model to get the result that has a high correct rate reaching to 90% with little human intervention. In this paper, the design of some algorithms is original. Combined with the present feasible algorithm, we get an ideal result.