(Translated from Press Article)
The development of Korean subtitles on domestic online video streaming services (aka OTT) is currently underway through research and development initiatives. With the support of the government and a specialized research management agency, efforts are being made to ensure accessibility for the media vulnerable population.
In April, the Ministry of Science and ICT and the ‘Korea Institute for Information Technology Planning and Evaluation’ (IITP), led by Director Seong-bae Jeon, launched the "Development of Colloquial Broadcasting Content Subtitle Production and Editing Technology" R&D project.
The main objective of this project is to utilize artificial intelligence (AI) to recognize audio content and automatically translate and generate subtitles while maintaining synchronization. The purpose is to enhance the "barrier-free" feature in the OTT sector, aiming to improve service accessibility for the elderly and people with disabilities. Platforms like Netflix provide closed captions that not only include dialogue but also sound cues and speaker identification for all their content.
The need to expand and implement such services in domestic OTT platforms has been emphasized for some time. In response to this demand, the government and IITP have collaborated to prepare for the necessary tasks.
With Twigfarm leading, Cystran, SBS, FurmoDT, and Wavve all bring in their strengths for synergy.
Cystran's technology focuses on improving the accuracy of subtitle generation by considering the genre of video content and the contextual information before and after the subtitle. Taking into account the differences in commonly usedwords in genres such as action or melodrama, they conduct additional training beyond the basic engine to maximize the accuracy of subtitle translation and generation. They also consider not only individual sentences but also the surrounding sentences to further enhance accuracy.
Twigfarm, particularly active in platform development, aims to bring together broadcasters, OTT operators, and subtitle companies on a single platform based on the cloud. They plan to collaborate with suitable AI models and allocate resources in stages. By automating labor-intensive tasks such as subtitle production and result verification, they anticipate significant time and cost savings. This will lead to a win-win situation for both users and service providers.
The generated subtitles using these technologies are not limited to the Korean language alone. Yu-seok Kim, CEO of Cystran, mentioned plans to expand the scope to include languages such as English, Chinese, Japanese, and eventually European and Southeast Asian languages, which could contribute to the global expansion of domestic OTT services.
The effectiveness of the project is of particular interest. Sun-ho Baek, CEO of Twigfarm, stated that “Through the efforts of the government and IITP, a smooth demonstration involving broadcasters, OTT operators, and other demand parties will be conducted”. He also revealed plans to launch a pilot service next year, “Aiming to enhance the convenience of OTT usage for all users”.
Additionally, there are already existing technologies supported by the government for the media vulnerable population. Korea’s ‘Electronics and Telecommunications Research Institute’ (ETRI) has developed a technology that translates Korean sentences into sign language and represents them using avatars, providing an "emotion expression service" to support visually impaired individuals in watching broadcasts. Han-kyu Lee, the R&D lead, mentioned ongoing efforts to commercialize the technology through collaboration with technology companies. He stated “With the acquisition of datasets and advanced modeling, it would be possible to apply the technology to OTT services as well.”