The Bidirectional Long-Short-Term Memory Neural Network based Word Retrieval for Arabic Documents
Journal Title: Transactions on Machine Learning and Artificial Intelligence - Year 2015, Vol 3, Issue 1
Abstract
The reflow from Arabic document image collections is a challenging task. This is partly due to the insolubility of the Arabic script. Because of the peculiarity of the whole body of the Arabic words, namely connectivity between the characters, thereby the segmentation of An Arabic word is very arduous and also the variability of the handwritten styles and shapes as well as deceleratation in the print therefore Arabic document repositories are not liable to indexation and reflow. In this paper, we are proposing an idea for reflow coherent Arabic document in response to an appropriate Arabic query word. We present a novel approach at reflow on an Arabic document image, using a Bidirectional Long-Short-Term Memory Neural Network. The designed to take relating to information into account, these networks can maintain Arabic word images that cannot be durable segmented into individual Arabic characters. The partitioning Arabic word, we easier the problem and receive elevated reflow rates. The proposed capable reflow scheme avoids unambiguous recognition of Arabic characters. An experimental evaluation on a dataset of Arabic word images conjunct from handwritten notebook show good precision even in the impendence of printing transformation and deceleratation. The reflow Arabic word performance is comparison with baseline methods. These results encourage the development of real world systems for word reflow for Arabic documents
Authors and Affiliations
Dr. Yusuf Perwej
Keywords
Related Articles
- EP ID EP278943
- DOI 10.14738/tmlai.31.863
- Views 111
- Downloads 0
Publishing Student Graduation Projects Based on the Semantic Web Technologies
Linked Open Data is already widely available in several industries, such as libraries, biomedicine and Linked government data, etc. for data sharing, promoting semantic web development and maintaining a global culture of...
Generic Platform to Generate Educational Components
In this research, we will present all the concepts necessary for the implementation of a warehouse of educational components. We define three classes of educational components: Learning components, services components an...
Face Spoofing and Counter-Spoofing: A Survey of State-of-the-art Algorithms
In the current scenario of biometric-based identity verification, a face is still being proved to be an essential physiological evidence for successful person identification without letting know the target. Nevertheless,...
Stress Management in Primary Caregivers: A Health Challenge
Technology based advances in healthcare are leading to social changes that, in turn, will require innovative responses from health systems. Mainly due to these advances, life expectancy grows so, the aging of world popul...
Using Artificial Intelligence to Predict Animal Behaviour in Food Webs
Overfishing of species in the marine life has caused oceans to become deserts at a fast pace. The population of specific species such as Cod and Haddock has reduced over the years. This has affected countries that hugely...