Skip to content

Latest commit

 

History

History
15 lines (12 loc) · 546 Bytes

File metadata and controls

15 lines (12 loc) · 546 Bytes

Text-NonText-Separation

In Brief: This is a Research Project(ML Based) under supervision of Associate Prof Ram Sarkar(Jadavpur University) Here we have proposed a novel text/non-text separation method in printed documents.In our Algorithm we have used Rotation invariant of Texture Based Local Ternary pattern and to eliminate the redundant features we have used Binary Particle Swarm Optimization Algorithm .

Classification Accuracy: Greater than 95%,given in the attached research paper

Language used: Matlab

Time span: sept-19 - Jan-20