Novel LVCSR Decoder Based on Perfect Hash Automata and Tuple Structures – SPREAD –
Journal Title: International Journal of Advanced Computer Science & Applications - Year 2014, Vol 5, Issue 5
Abstract
The paper presents the novel design of a one-pass large vocabulary continuous-speech recognition decoder engine, named SPREAD. The decoder is based on a time-synchronous beam-search approach, including statically expanded cross-word triphone contexts. An approach using efficient tuple structures is proposed for the construction of the complete search-network. The foremost benefits are the important space savings and higher processing speed, and the compact and reduced size of the tuple structure, especially when exploiting the structure of the key. In this way, the time needed to load the ASR search-network into the memory is also significantly reduced. Further, the paper proposes and presents the complete methodology for compiling general ASR knowledge sources into a tuple structures. Additionally, the beam search is enhanced with the novel implementation of a bigram language model Look-Ahead technique, by using tuple structures and a caching scheme. The SPREAD LVCSR decoder is based on a token-passing algorithm, capable of restricting its search-space by several types of token pruning. By using the presented language model Look-Ahead technique, it is possible to increase the number of tokens that can be pruned without decoding precision loss.
Authors and Affiliations
Matej Rojc, Kacic Zdravko
Keywords
Related Articles
- EP ID EP105025
- DOI 10.14569/IJACSA.2014.050504
- Views 102
- Downloads 0
Junction Point Detection and Identification of Broken Character in Touching Arabic Handwritten Text using Overlapping Set Theory
Touching characters are formed when two or more characters share the same space with each other. Therefore, segmentation of these touching character is very challenging research topic especially for handwritten Arabic de...
A Classification Model for Imbalanced Medical Data based on PCA and Farther Distance based Synthetic Minority Oversampling Technique
Medical data are extensively used in the diagnosis of human health. So it has played a vital role for physicians as well as in medical engineering. Accordingly, many types of research are going on related to this to have...
English-Arabic Hybrid Machine Translation System using EBMT and Translation Memory
The availability of a machine translation to translate from English-to-Arabic with high accuracy is not available because of the difficult morphology of the Arabic Language. A hybrid machine translation system between Ex...
The Role of Hyperspectral Imaging: A Literature Review
Optical analysis techniques are used recently to detect and identify the objects from a large scale of images. Hyperspectral imaging technique is also one of them. Vision of human eye is based on three basic color (red,...
Development of Copeland Score Methods for Determine Group Decisions
Voting method requires to determine group decision of decision by each decision maker in group. Determination of decisions by group of decision maker requires voting methods. Copeland score is one of voting method that h...