Automatic Ship Types Classification in Silhouette Images
M. R. Noroozi1, A. Ramezani2, M. Aghababaee3
1M. R. Noroozi, Electrical & Electronics Engineering Department, Imam Khomeini Maritime Sciences University, Noshahr, Iran.
2A. Ramezani, Electrical & Electronics Engineering Department, Imam Khomeini Maritime Sciences University, Noshahr, Iran.
3M. Aghababaee, Electrical & Electronics Engineering Department, Imam Khomeini Maritime Sciences University, Noshahr, Iran.
Manuscript received on September 24, 2014. | Revised Manuscript received on October 03, 2014. | Manuscript published on October 30, 2014. | PP: 52-56  | Volume-4 Issue-1, October 2014. | Retrieval Number:  F3389083614/2013©BEIESP

Open Access | Ethics and Policies | Cite
© 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: Object identification or object classification is an important task in computer vision and pattern recognition. Silhouette image comprises many features which can be used for these demands. In this paper Discrete Hartley Transform (DHT) and Discrete Cosine Transform (DCT) are used for feature extraction from silhouette image. These features are then applied to the neural network for ship type classification. Ship features from different view (only 4 features in each image) were trained with feed forward back propagation neural network and accuracy was satisfied for testing over 50 images, also this algorithm is stands up robustly against the noise and can be used for classification another things such as animals, people , vehicles, etc.
Keywords: Pattern recognition, Object classification silhouette image, DHT (Discrete Hartley Transform), DCT, ship type classification.