Writer Identification with Hybrid Model using Hybrid HMM and ANN
Vinita Balbhim Patil1, Rajendra R Patil2
1Vinita Balbhim Patil, Assistant Professor, Department of Electronics and Communication Engineering.
2Dr. Rajendra R Patil, Professor & HOD in Dept.of ECE, GSSSIETW, Mysuru, India.
Manuscript received on 6 August 2019. | Revised Manuscript received on 11 August 2019. | Manuscript published on 30 September 2019. | PP: 1656-1661 | Volume-8 Issue-3 September 2019 | Retrieval Number: C4435098319/19©BEIESP | DOI: 10.35940/ijrte.C4435.098319
Open Access | Ethics and Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: In this paper, writer identification is performed with three models, namely, HMMBW, HMMMLP and HMMCNN. The features are extracted from the HMM and are classified using Baum Welch algorithm (BW), Multi layer perceptron (MLP) model and Convolutional neural network (CNN) model. A dataset, namely, VTU-WRITER dataset is created for the experiential purpose and the performance of the models were tested. The test train ratio was varied to derive its relation to accuracy. Also the number of states was varied to determine the optimum number of states to be considered in the HMM model. Finally the performance of all the three models is compared.
Keywords: Convolutional neural network (CNN) model ,Hidden markov models, Multi layer perceptron (MLP) model , Writer identification.
Scope of the Article: Probabilistic Models and Methods.