IMPROVING FEATURE MAPS IN EARLY LAYERS OF CONVOLUTIONAL NEURAL NETWORKS USING OTSU METHOD
Al-furas, A., AL-dosuky, M., Hamza, T. (2016). IMPROVING FEATURE MAPS IN EARLY LAYERS OF CONVOLUTIONAL NEURAL NETWORKS USING OTSU METHOD. EKB Journal Management System, 16(2), 37-45. doi: 10.21608/ijicis.2018.10905
A Al-furas; M AL-dosuky; Taher Hamza. "IMPROVING FEATURE MAPS IN EARLY LAYERS OF CONVOLUTIONAL NEURAL NETWORKS USING OTSU METHOD". EKB Journal Management System, 16, 2, 2016, 37-45. doi: 10.21608/ijicis.2018.10905
Al-furas, A., AL-dosuky, M., Hamza, T. (2016). 'IMPROVING FEATURE MAPS IN EARLY LAYERS OF CONVOLUTIONAL NEURAL NETWORKS USING OTSU METHOD', EKB Journal Management System, 16(2), pp. 37-45. doi: 10.21608/ijicis.2018.10905
Al-furas, A., AL-dosuky, M., Hamza, T. IMPROVING FEATURE MAPS IN EARLY LAYERS OF CONVOLUTIONAL NEURAL NETWORKS USING OTSU METHOD. EKB Journal Management System, 2016; 16(2): 37-45. doi: 10.21608/ijicis.2018.10905

IMPROVING FEATURE MAPS IN EARLY LAYERS OF CONVOLUTIONAL NEURAL NETWORKS USING OTSU METHOD

International Journal of Intelligent Computing and Information Sciences
Article 3, Volume 16, Issue 2, April 2016, Page 37-45  XML PDF (882.9 K)
Document Type: Original Article
DOI: 10.21608/ijicis.2018.10905
View on SCiNiTO View on SCiNiTO
Authors
A Al-furas1; M AL-dosuky1; Taher Hamza2
1Faculty of Computer and Information,Mansoura University, Egypt.
2Computer Science Department Faculty of Computer and Information Sciences, Mansoura University - Egypt
Abstract
Abstract: A novel deep architecture Thresholding Convolution Neural Network (ThCNN) progresses in this paper; Which is a simple and effective method to regularizing features map in the early layers of Convolution Neural Network(CNN). One of the issues identified with deep learning is the features in early layers that robustness and discriminativeness. In this paper, we compute the optimal global threshold to determine the features that are passed to the next layers. We then evaluate ThCNN on an MNIST dataset comparing it CNN by applying multiple trained models. It yield decent accuracy compared to traditional CNN. It gives a 99.5%
Statistics
Article View: 170
PDF Download: 394