Text and non-text segmentation for automatic interpretation of hand-drawn diagrams | |
| Author | Buntita Pravalpruk |
| Call Number | AIT Diss. no.CS-24-02 |
| Subject(s) | Data mining Deep learning (Machine learning) |
| Note | A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy in Computer Science |
| Publisher | Asian Institute of Technology |
| Abstract | Handwriting is a natural way to communicate and exchange ideas, but converting hand written diagrams to application-specific digital formats requires skill and time. Auto matic offline handwritten document conversion can save time, but diagrams and text require different recognition engines. Since accurate segmentation of handwritten diagrams can improve the accuracy of later diagram recognition steps, I propose several methods in two categories of strategies to classify text and objects in diagrams using local features of image contours or end-to-end using a deep learning method called DeepDP. I experiment the models on two different dataset; OHFD and the new dataset BPMN,and I compare the performance of those strategies. Deep learning method is the best methods achieve an accuracy of 98.8% on those two dataset. |
| Year | 2024 |
| Type | Dissertation |
| School | School of Engineering and Technology |
| Department | Department of Information and Communications Technologies (DICT) |
| Academic Program/FoS | Computer Science (CS) |
| Chairperson(s) | Dailey, Mathew N. |
| Examination Committee(s) | Mongkol Ekpanyapong;Sanparith Marukatat |
| Degree | Thesis (Ph.D.) - Asian Institute of Technology, 2024 |