RECONSTRUCTION OF TORN DOCUMENTS USING SUPPORT VECTOR MACHINE (SVM) METHOD
Keywords:
Torn pieces or documents; Image processing; Support Vector Machine (SVM) method; Edge detection; Using in forensic science; Do core recovery algorithm; Documents reconstruction; Reconstruction techniques; Document image processingAbstract
The main task of my thesis is to reconstruct the torn pieces of any paper. This paper can be a document
or an image. Documents could be shredded for any reason such as, one may do it intentionally and it could happen
by mistake also but the main disadvantage is that, the information will be lost which are for certain cases extremely
important. Reconstruction of a torn document can be very beneficial for people from various fields, such as in
defence, investigative science, archaeology and forensics fields.
Torn documents can be hand shredded or manually shredded. One can shred it vertically or
horizontally. The task will be solved by using Support Vector M a c h i n e (SVM) method. Torn pieces are
combined by comparing edge length, angles, page margin and by matching the colors of images or text documents.
Therefore, my task is divided in three stages: read the image, find edges and fragmentation of the edges. In
addition to that, SVM is a newly invented method which is also suitable for the documents which have high
resolution.
