Combining RtL and LtR HMMs to recognise handwritten Farsi words of small‐ and medium‐sized vocabularies
In this study, a method for holistic recognition of handwritten Farsi words is proposed, which fuses the outputs of right‐to‐left (RtL) and left‐to‐right (LtR) hidden Markov models (HMMs). The experimental results on 16,000 images of 200 names of Iranian cities, from the ‘Iranshahr 3’ are presented...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Wiley
2018-09-01
|
| Series: | IET Computer Vision |
| Subjects: | |
| Online Access: | https://doi.org/10.1049/iet-cvi.2017.0645 |
| _version_ | 1797684850405146624 |
|---|---|
| author | Seyed Ali Asghar Abbaszadeh Arani Ehsanollah Kabir Reza Ebrahimpour |
| author_facet | Seyed Ali Asghar Abbaszadeh Arani Ehsanollah Kabir Reza Ebrahimpour |
| author_sort | Seyed Ali Asghar Abbaszadeh Arani |
| collection | DOAJ |
| description | In this study, a method for holistic recognition of handwritten Farsi words is proposed, which fuses the outputs of right‐to‐left (RtL) and left‐to‐right (LtR) hidden Markov models (HMMs). The experimental results on 16,000 images of 200 names of Iranian cities, from the ‘Iranshahr 3’ are presented and compared with those methods using only RtL or LtR models. Experimental results show that the main sources of error are similar beginnings or similar endings of the words. Since RtL and LtR models when dealing with the words behave differently, there is notable error diversity between the two classifiers in such a way that their combination increases the recognition rate. Compared to the RtL‐HMM, the product of output scores of the RtL and LtR‐HMMs reduces the classification error to about 6, 6 and 3%, for three different feature sets. A subjective error analysis on the results is also provided. |
| first_indexed | 2024-03-12T00:35:40Z |
| format | Article |
| id | doaj.art-fe3296074f334ad9bd58b6a220bcb8a7 |
| institution | Directory Open Access Journal |
| issn | 1751-9632 1751-9640 |
| language | English |
| last_indexed | 2024-03-12T00:35:40Z |
| publishDate | 2018-09-01 |
| publisher | Wiley |
| record_format | Article |
| series | IET Computer Vision |
| spelling | doaj.art-fe3296074f334ad9bd58b6a220bcb8a72023-09-15T09:46:18ZengWileyIET Computer Vision1751-96321751-96402018-09-0112692593210.1049/iet-cvi.2017.0645Combining RtL and LtR HMMs to recognise handwritten Farsi words of small‐ and medium‐sized vocabulariesSeyed Ali Asghar Abbaszadeh Arani0Ehsanollah Kabir1Reza Ebrahimpour2Department of Electrical and Computer EngineeringTarbiat Modares UniversityAl Ahmad StreetIranDepartment of Electrical and Computer EngineeringTarbiat Modares UniversityAl Ahmad StreetIranFaculty of Computer EngineeringShahid Rajaee Teacher Training UniversityTehranIranIn this study, a method for holistic recognition of handwritten Farsi words is proposed, which fuses the outputs of right‐to‐left (RtL) and left‐to‐right (LtR) hidden Markov models (HMMs). The experimental results on 16,000 images of 200 names of Iranian cities, from the ‘Iranshahr 3’ are presented and compared with those methods using only RtL or LtR models. Experimental results show that the main sources of error are similar beginnings or similar endings of the words. Since RtL and LtR models when dealing with the words behave differently, there is notable error diversity between the two classifiers in such a way that their combination increases the recognition rate. Compared to the RtL‐HMM, the product of output scores of the RtL and LtR‐HMMs reduces the classification error to about 6, 6 and 3%, for three different feature sets. A subjective error analysis on the results is also provided.https://doi.org/10.1049/iet-cvi.2017.0645combined RtL-LtR HMMshandwritten Farsi word recognitionmedium-sized vocabulariessmall-sized vocabulariesleft-to-right hidden Markov modelsright-to-left hidden Markov models |
| spellingShingle | Seyed Ali Asghar Abbaszadeh Arani Ehsanollah Kabir Reza Ebrahimpour Combining RtL and LtR HMMs to recognise handwritten Farsi words of small‐ and medium‐sized vocabularies IET Computer Vision combined RtL-LtR HMMs handwritten Farsi word recognition medium-sized vocabularies small-sized vocabularies left-to-right hidden Markov models right-to-left hidden Markov models |
| title | Combining RtL and LtR HMMs to recognise handwritten Farsi words of small‐ and medium‐sized vocabularies |
| title_full | Combining RtL and LtR HMMs to recognise handwritten Farsi words of small‐ and medium‐sized vocabularies |
| title_fullStr | Combining RtL and LtR HMMs to recognise handwritten Farsi words of small‐ and medium‐sized vocabularies |
| title_full_unstemmed | Combining RtL and LtR HMMs to recognise handwritten Farsi words of small‐ and medium‐sized vocabularies |
| title_short | Combining RtL and LtR HMMs to recognise handwritten Farsi words of small‐ and medium‐sized vocabularies |
| title_sort | combining rtl and ltr hmms to recognise handwritten farsi words of small and medium sized vocabularies |
| topic | combined RtL-LtR HMMs handwritten Farsi word recognition medium-sized vocabularies small-sized vocabularies left-to-right hidden Markov models right-to-left hidden Markov models |
| url | https://doi.org/10.1049/iet-cvi.2017.0645 |
| work_keys_str_mv | AT seyedaliasgharabbaszadeharani combiningrtlandltrhmmstorecognisehandwrittenfarsiwordsofsmallandmediumsizedvocabularies AT ehsanollahkabir combiningrtlandltrhmmstorecognisehandwrittenfarsiwordsofsmallandmediumsizedvocabularies AT rezaebrahimpour combiningrtlandltrhmmstorecognisehandwrittenfarsiwordsofsmallandmediumsizedvocabularies |