To search, Click below search items.

 

All Published Papers Search Service

Title

Applications of Meta-Heuristic Algorithm with Back Propagation Classifier for Handling Class of General Fish Models

Author

Bushra S. Al Smadi

Citation

Vol. 16  No. 10  pp. 38-45

Abstract

Pattern recognition is a system that recognizes isolated patterns of interest that could be an image. Many models such as noise, distortion, overlap, and errors in the segmentation results and obstruction of image’s objects could occur during the process of image recognition. The aim of this study is to develop a system to recognize isolated fish object in the image based on a combination of significant extracted features using anchor points, texture and statistical measurements. A generic fish classification method could be performed using a hybrid meta-heuristic algorithms (genetic algorithm with Great Deluge (GD) algorithm) with back-propagation algorithm (GAGD-BPC). Thus, it is used to classify the images of fish into dangerous and non-dangerous families, to recognize the dangerous fish families into Predatory and Poison fish family, and to recognize the non-dangerous fish families into garden and food fish family. A prototype to deal with the problem of fish images classification is presented in this research work. The proposed prototype has been tested based on 24 fish families, each family contains different number of species. Therefore, it has performed the classification process successfully. The experimental tests have been performed based on 320 distinct fish images that were divided into 220 images for training phase and 100 images for testing phase. An overall accuracy recognition rate was 83.2%, which was obtained using the proposed GAGD-BPC.

Keywords

Great deluge (GD) algorithm, Feature Extraction, Back Propagation Classifier, Anchor Measurements and Genetic Algorithm.

URL

http://paper.ijcsns.org/07_book/201610/20161007.pdf