Achieving higher accuracy than ever before, Science Arts, Inc. has adopted the speech recognition API provided by Advanced Media Corporation's AmiVoice Cloud Platform as the default speech recognition engine for Buddycom's speech-to-text functionality.
In addition to voice, live video, and location-based team communication (IoT), Buddycom also provides AI-based digital assistant services. It has also been introduced as a new communication tool to support the workplace in all kinds of industries, including railways, airlines, distribution, caregiving, and factories. The ‘speech-to-text’ function is very useful. It enables automatic transcription of the speech used in Buddycom.
AmiVoice, is a highly accurate speech recognition engine with No. 1 market share*, strong in Japanese and technical terms.
Buddycom’s ‘speech-to-text’ functionality has become even easier to use by connecting Buddycom to the ‘AmiVoice Cloud Platform’s’ speech recognition API.
In addition to improving the accuracy of Japanese recognition, it also improves the accuracy of recognition of technical terms, idiosyncratic phrases, and dialects. Punctuation and question marks are automatically added, and fillers such as ‘uh,’ ‘that,’ and ‘um’ are automatically removed. The result is more accurate and easier to understand writing. This means that words spoken in Buddycom will appear as more readable sentences than before. This makes team communication in the field using Buddycom smoother than ever.
To further improve speech recognition accuracy, Buddycom users will be able to select industry-specific engines such as medical and nursing. The addition of proper names such as product names, industry terms, and people’s names to the lexicon further strengthens the cooperation, with the goal of improving the recognition accuracy of words that are difficult to accurately convert to text with ordinary speech recognition engines.
Buddycom will continue to promote on-site DX in all industries and sectors with the ‘AmiVoice Cloud Platform’.
[The Buddycom and AmiVoice Cloud Platform Collaboration Brings New Opportunities].
●The No.1 speech recognition engine in the market improves Japanese recognition accuracy
The accuracy of Buddycom’s ‘speech-to-text’ function is improved by linking it with AmiVoice, a high-precision speech recognition engine that has the No. 1 share※ in the speech recognition market and is strong in Japanese and technical terminology.
●Automatic Deletion of, ‘uh,’ ‘Um,’ Fillers
Fillers such as ‘um,’ ‘and ah…,’and ‘um’ will be automatically removed. Punctuation and question marks are also automatically added. This helps create more accurate and easier to understand phrases.
●Business-specific Language Filtering
Uses a language model that omits inappropriate or improper non-business terms.
By eliminating unnecessary terms, it reduces the potential for misinterpretation, making it ideal for business use.
Register words for recognition, such as product names and proper nouns.
This improves recognition of specific words and phrases, such as company jargon. (The word registration function is available in Japanese only.)
A ‘domain-specific engine’ that specializes in technical and industry-specific terms, such as medical and nursing, can be selected in addition to the ‘general-purpose engine’ that can be used in various scenarios and businesses to improve recognition accuracy.
●Individual Tuning of the Speech Recognition Engine
Custom tuning is performed to improve accuracy for conversations that cannot be handled by word registration or industry-specific engines.
What is the AmiVoice Cloud Platform?
To help you to implement speech recognition functionality without embedding special libraries in client applications, we provide a speech recognition API service. This significantly reduces the time required for traditional speech recognition development. It also enables easy integration of a speech recognition engine with high recognition accuracy. AmiVoice is a speech recognition engine that has the No. 1 market share in the speech recognition market and is strong in the spoken language area. It can be used for speech characterization, speech dialog, speech control, speech translation, applications, websites, and more.