Three new apps to enhance digital experience in Bangla
The government has developed three AI-based software promoting the Bengali language in the digital space, enabling users to convert Bengali text to speech and speech to text and extract text from images.
The Bangladesh Computer Council, under the "Enrichment of Bengali Language in Information Technology through Research and Development" (EBLICT) project, will release the software for users tomorrow on the occasion of the International Mother Language Day.
There will also be a font opened for users that day, digitally representing the Bengali characters, symbols and glyphs.
"These tools are unique in Bangladesh," said Mahbub Karim, project director of EBLICT.
Each of the components in the software is developed using artificial intelligence and developed by the local engineers.
Although Google has similar software, these are not free, he said, adding that they have plans to make these software feasible, sustainable and resourceful in future.
The Bengali text-to-speech software, named Uccharon, will convert text into spoken words. As the application reads text from documents, websites and screen windows, it will highly benefit visually impaired people too.
Developed with both male and female voices, the software is accessible at read.bangla.gov.bd from today.
The Bengali speech-to-text software is named "Kotha". Accessible via voice.bangla.gov.bd, it transforms clear and properly spoken Bengali words into text in clear environments. Future updates of the software will include major Bengali punctuation marks. Android mobile users will also be able to use the project's keyboard app.
The Optical Character Recognition software, Borno, facilitates the conversion of Bengali uneditable text into editable text, particularly in PDF or JPEG files. Its OCR feature renders Bengali text into a composed format.
Borno OCR also recognises various elements like tables, common English words, institution names and logos in government documents. Accessible via ocr.bangla.gov.bd, it can be used from any browser.
The Bangla Unicode font "Purno" addresses font-related challenges, encompassing essential typographic elements for Bengali publications.
According to the BCC, it can accurately represent Bengali linguistic nuances and suit both institutional and digital platforms.
"Despite numerous Bengali fonts, the absence of a standard one usually creates glyph set inconsistencies. To address this, we crafted a unified font suitable for public, private, print, publishing and web platforms," said Mamun Or Rashid, a consultant on the project.
Unique features of this font include comprehensive glyphs, consistent visual form, varied allographs and balanced sizing, ensuring seamless representation of Bengali across mediums, said Rashid, who is also an assistant professor at the Jahangirnagar University's department of Bangla.
Meanwhile, under the project, the BCC has initiated efforts to preserve and digitise 40 ethnic languages, including 14 endangered ones.
Already 12,000 minutes of predefined and spontaneous data have been collected with an average of 300 minutes per language, said Rashad Kabir, managing director of Dream71 Bangladesh, which partnered with the BCC for the initiative.
Their team went to the remotest part of Bangladesh, including the three hill tracts, Sylhet, Sreemangal, Rajshahi, Dinajpur and the Mymensingh region to collect data on the dialects.
Fonts and keyboards of six languages have also been developed along with a language portal to make the digitised resources accessible to a global audience.
"These portals allow users to learn, research and appreciate the diverse ethnic linguistic Bangladesh," Kabir added.
Comments