Homechevron_rightNewschevron_rightNews Releasechevron_rightAI voice recognition transcription service “ProVoXT” is equipped with an engine for medical conferences and Bi-LSTM

AI speech recognition transcription service "ProVoXT" equipped with medical conference engine and Bi-LSTM

Significantly improved recognition rate with dedicated engine and deep learning technology

Advanced Media Co., Ltd. (Headquarters: Toshima-ku, Tokyo; Chairman and President: Kiyoyuki Suzuki; hereinafter referred to as "Advanced Media") has added a new engine for medical conferences to its speech recognition cloud-based transcription service "ProVoXT." , is equipped with bidirectional deep learning technology "Bi-LSTM (Bidirectional Long Short-Term Memory)".


【background】


At medical institutions, many meetings are held every day, such as treatment policy, intra-hospital collaboration, medical safety promotion meetings, verification meetings, management meetings, etc., and minutes are created accordingly. Minutes are indispensable for decision-making, policy recording, information sharing, etc., but they are also a factor in labor shortages and chronic overtime, and improving the efficiency of transcription is a priority for each organization. It has become one.


[Equipped with an engine for medical conferences]


"ProVoXT" is a cloud-based transcription service that uses voice recognition to semi-automate the creation of minutes, quickly and efficiently. Since it can be operated without any limit on the number of users for a monthly fee, it promotes work style reform throughout the institution and realizes improvements in operational efficiency and productivity.

In addition to the regular conferencing engine, we have now added a voice recognition engine for medical conferences. Because it uses an engine specialized for medical terminology, technical terms such as "droplet infection," "acute infarction," and "lymph node enlargement" can be transcribed smoothly.


[Implementation of Bi-LSTM technology]

"Bi-LSTM" performs processing from a bidirectional time axis by predicting and adding not only past but also future information. Compared to the recurrent neural network technology ``LSTM (Long Short-Term Memory)'' implemented in the conventional speech recognition engine AmiVoice, we achieved an error improvement rate of up to 25%, significantly improving the recognition rate. It is implemented in the engine installed in ProVoXT.


■ProVoXT, details


http://www.advanced-media.co.jp/products/service/provoxt

At Advanced Media, we will continue to focus on providing new solutions and customer services that contribute to operational efficiency in a wide range of industries and business categories.

Above

Inquiries regarding this matter

Management Promotion Headquarters Public Relations Team

TEL:03-5958-1307
FAX: 03-5958-1032

Inquiries regarding the contents of this report

VoXT Division

TEL:03-5958-1035
FAX: 03-5958-1033

Inquiries about this product

Japan's No.1 in Market ShareJapan's No.1 in Market ShareAmiVoiceⓇAmiVoiceⓇ

*Source: ecarlate LLC "Voice Recognition Market Trends 2025"
Speech recognition software/cloud service market

Write with your voice, move with your voice.
AI voice recognition AmiVoice
In various business situations,
This is a technology that enables natural communication between people and machines.