A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL.
The file type is application/pdf.
Filters
Application Of Smooth Ergodic Hidden Markov Model In Text To Speech Systems
2008
Zenodo
Preserved Fulltext
The proposed method is a probabilistic model and is based on Smooth Ergodic Hidden Markov Model. This model can be considered as an extension to HMM. ...
In developing a text-to-speech system, it is well known that the accuracy of information extracted from a text is crucial to produce high quality synthesized speech. ...
it can be used to address the problem with the HMM. ...
doi:10.5281/zenodo.1070939
fatcat:5jkpbtyxprcsvavp4i3v73zemy
Modeling and training options for handwritten Arabic text recognition
2017
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2018; you can also visit the original URL.
The file type is application/pdf.
This leads to a significantly compact text recognizer as compared to the HMM systems that use the standard contextual models. ...
for Arabic text. ...
This, in turn, is used to retrain a new HMM system which is finally used for text recognition.Finally, the third approach is to use neural networks with HMMs in a hybrid way where the state output probability ...
doi:10.17877/de290r-17923
fatcat:6phbwzcoffbvdl4lo3novlxtw4
Unconstrained recognition of offline Arabic handwriting using generative and discriminative classification models
[article]
2018
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL.
The file type is application/pdf.
The developed system is used to create more than 12000 synthesized samples that have been added to the database. ...
A letter frequency has analysis showed that the database iii iv exhibits letter frequencies very similar to that of large corpora of digital text, which proves the database usefulness. ...
The best recognition rate achieved is 90%, when a combination of neural networks is chosen. ...
doi:10.25673/4590
fatcat:ewgowrv6abg6nldscxgt6kdoeu
Multi-Script Off-Line Signature Verification
2018
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL.
The file type is application/pdf.
A multi-lingual country like India has many different scripts that are used for writing as well as for signing purposes based on different locations or regions. ...
In India, a single official transaction sometimes needs signatures using more than one script. ...
A technique for the automatic recognition of Arabic printed text using artificial neural networks was presented by Sarfraz et al. [300] . ...
doi:10.25904/1912/2627
fatcat:hinl3cjsnvbuhd7ryipre2p7eu