Voice2Med Speech Recognition System in Medicine

The main articles are:

• Physician speech recognition
• Voice biometrics
• Biometric identification technologies

2023: Availability for doctors of nine specialties

On May 31, 2023, the Group of Companies CST presented an updated product release of the Voice2Med - voice filling out medical protocols based on. artificial intelligence Now the possibilities of Voice2Med have been significantly expanded - to the available dictionaries of a general practitioner and pediatrician,,, radiologist CT MRI pathomorphologist, diagnostic doctor ultrasonic and endoscopist, surgeon stomatologist cardiologist, and a dictionary of medical terms - a dictionary has been added. The solution oncologist is included in, register of domestic software supports work on Russian. operating systems

It is enough for a doctor to pronounce the result of a study or examination into a headset, Voice2Med converts speech into text and provides a completed protocol that requires a minimum of adjustments.

We announce the availability of voice input for doctors in nine specialties. Now even more specialists will be able to appreciate the advantages of speech recognition and optimize work with documents. At the same time, each specialized medical dictionary is constantly developing, for this we have implemented a function with which each doctor can independently add new terms and expressions. In addition, by translating the Voice2Med to new generation speech recognition, we managed to significantly reduce the requirements for headsets, which significantly expands the ability to scale voice input, said Dmitry Dyrmovsky, CEO of the CST Group of Companies.

The first large-scale implementation of the Voice2Med was implemented in Moscow, which became possible thanks to a single digital healthcare platform developed by the capital's social development complex and the Department of Information Technologies. The product is already successfully used by doctors of radiation diagnostics: it works in 120 medical institutions in Moscow. The Center for Diagnostics and Telemedicine recorded in medical organizations of the Moscow Department of Health a reduction in more than 20% of the time for preparing medical protocols.

2022

Filling more than 210 thousand medical protocols of radiation studies using the voice input service

Moscow policlinics In and hospitals voice , more than 210 thousand medical radiation research protocols were filled out using the input service. The technology speech recognition allows you to dictate the description of the study and save time doctors on information manual input, which speeds up this work by more than 20%. This was announced on January 9, 2022 by the deputy head. Department of Health of Moscow Ilya Tyrov

As of January 20223, the solution is used in children's and adult clinics, the Moscow reference center, multidisciplinary hospitals and medical organizations providing specialized assistance. Radiologists by voice more than 210 thousand medical protocols have already been filled out in the capital. The number of documents filled out using the digital assistant is growing; since the beginning of 2022, 80% more protocols have been filled out than in 2021. At the same time, the quality of speech recognition reached 98%. Speech recognition technology allows you to dictate the description of the study, and not enter information manually, which accelerates the preparation of research protocols by more than 20%, "said health care Moscow Ilya Tyrov, Deputy Head of the Department.

Thanks to the unified digital healthcare platform developed by the Social Development Complex of the city Moscow and, Department of Information Technology it was possible to implement the first large-scale use of voice input in the country. The product is based on a solution artificial intelligence that converts the doctor's voice into text. During the interpretation of the study, the doctor dictates information that is instantly decrypted and transferred to an open protocol medical information system in the form of text.

We are working on the development of voice input together with the solution developer - the CST group - from 2020. The feedback of the doctors of the Center for Diagnostics and Telemedicine allows developers to systematically and quickly replenish the dictionary with complex and rare medical terms, adapt to the pronunciation features of different doctors, and cover more and more new diagnostic and therapeutic tasks. And this is really a large-scale activity. During the work, various headsets were tested, the developers implemented integrations with several types of microphones. They also took care of the adaptation of doctors: they prepared educational and methodological materials, memos, instructions, video tutorials, - said Yuri Vasiliev, chief freelance specialist in radiation and instrumental diagnostics in Moscow, director of the Moscow Center for Diagnostics and Telemedicine.

The joint work made it possible to achieve the quality of speech recognition of 98%, including specialized medical vocabulary. We strive to transfer the experience of metropolitan radiologists, open up the possibilities of technological services for doctors of different specialties and have already made voice input available to doctors of eight specializations. Voice2Med is a completely domestic solution, does not contain borrowed components, is included in the register of domestic software, supports work on Russian operational systems, - said Dmitry Dyrmovsky, General Director of the CST Group of Companies.

Prize of the Government of the Russian Federation for the development of voice input for radiologists

Scientists of the Center for Diagnostics and Telemedicine of the DZM and the CST group of companies received the Prize of the Government of the Russian Federation for the adaptation and introduction of speech recognition technology for the preparation of protocols for X-ray and ultrasound studies. This was announced on December 26, 2022 by Zdrav.Expert representatives of the Center for Diagnostics and Telemedicine of DZM.

На фото (слева-направо): Mikhail Mishustin, Prime Minister of the Russian Federation, and Nikita Kudryavtsev, Junior Researcher, Standardization and Quality Control Sector, DZ CDT Award of
the Government of the Russian Federation
Photo: Youtube Channel of the Government of the Russian Federation

The technology was based on the Voice2Med solution - an artificial intelligence-based product for filling out medical protocols. The service allows radiologists to dictate the description of the study and not waste time entering information manually, this speeds up paperwork by more than 20%.

As part of the team that took part in the project:

• Alexey Petryaykin, Candidate of Medical Sciences, Associate Professor, Leading Researcher in the Standardization and Quality Control Sector of TsD & T DZM, Scientific Director of the Author's Team,
• Nikita Kudryavtsev, Associate Researcher, Standardization and Quality Control Sector, TsD & T DZM,
• Dmitry Semenov, Head of the Standardization and Quality Control Sector of TsD & T DZM,
• Daria Sharova, Head of Innovative Technologies at TsDiT DZM
• and researcher at the CST Group of Companies Maria Mitrofanova.

According to Yuri Vasiliev, chief freelance specialist in radiation and instrumental diagnostics in Moscow, director of the Moscow Center for Diagnostics and Telemedicine, Moscow radiologists are already using voice input technology in clinics and hospitals. As of the end of December 2022, hundreds of thousands of medical protocols for radiation research have already been filled out using the voice input service.

"Speech recognition technology allows doctors not to waste time entering information manually, radiologists dictate the description into the microphone, and the text immediately appears on the screen - this speeds up paperwork by more than 20%," said Yuri Vasiliev.

According to Dmitry Dyrmovsky, Director General of the CST Group of Companies, the first large-scale use of voice input was implemented in Moscow, which became possible thanks to a single digital healthcare platform developed by the capital's social development complex and the Department of Information Technologies.

"The high assessment of the decision and the receipt of the Prize of the Government of the Russian Federation in the field of science and technology is a great honor for us. The joint work made it possible to achieve the quality of speech recognition of 98%, including specialized medical vocabulary. We strive to transfer the unique experience of the capital's radiologists, open up the possibilities of modern technological services for doctors of different specialties and have already made voice input available to doctors of eight specializations, "added Dmitry Dyrmovsky.

As he stressed, Voice2Med is a completely domestic solution that does not contain borrowed components, is included in the register of domestic software, and supports work on Russian operational systems.

Filling out 160 thousand medical documents using voice input technology

Moscow continues to develop the services of a single digital healthcare platform. Voice input technology has proven itself and is in demand among radiation diagnostics specialists: over 160 thousand medical documents were filled out in two years of active use, and in the first half of 2022, 80% more protocols than in 2021. This was announced on July 8, 2022 by the CST company. This service is available to Moscow doctors and saves more than 20% of the time, said Yuri Vasiliev, director of the Center for Diagnostics and Telemedicine of the Moscow City Health Department.

The voice assistant Voice2Med created on the basis of artificial intelligence and allows you to fill out medical documents by voice. When examining patients or conducting a study, the doctor uses a headset to dictate information that is decrypted and automatically transferred to the open protocol of the medical information system. Thanks to operational feedback from doctors, the dictionary of the digital assistant is enriched with complex and rare medical terms, abbreviations and abbreviations, adapts to the pronunciation features of different doctors. Voice filling of medical documentation saves doctors time, improves the quality of protocol filling, improves the ergonomics of worksites. The voice interface is natural and intuitive, it is easy to study and master. The best practice of Moscow radiologists is being adopted in the regions - Voice2Med is working for July 2022 already in 65 regions of the country, new dictionaries are being developed - doctors of eight specialties can use voice input, the decision is being scaled, - commented Dmitry Dyrmovsky, General Director of the CST group of companies.

The voice assistant is constantly completed: it replenishes the dictionary with complex and rare medical terms, adjusts to the pronunciation characteristics of different doctors, covers new diagnostic and therapeutic tasks. Thanks to this, the accuracy of speech recognition reached 98%.

The first large-scale introduction of a stationary version of the voice assistant was implemented in Moscow, which became possible thanks to a single digital healthcare platform developed by the Moscow City Social Development Complex and the Department of Information Technologies.

Voice input for doctors of eight specialties and mobile workplace

On June 1, 2022, the CST Group of Companies presented the next product release of the Voice2Med with the possibility of voice filling of medical protocols based on artificial intelligence.

"The possibilities of Voice2Med are significantly expanded - to the available dictionaries of the general practitioner and pediatrician,, radiologist CT, MRI pathomorphologist, doctor ultrasound diagnostics and endoscopist cardiologist surgeon and added a dictionary of medical terms - a dictionary. stomatologist Now doctors of eight specialties can assess the possibilities of voice input. In addition, we have developed a mobile doctor's workplace made it possible to use voice input using, which tablet is convenient when a full-fledged workplace, with a stationary one, is not computer available. Mobile jobs can be in demand by ambulance specialists, pathomorphologists., Doctors forensic scientists on rounds. The solution allows you to fix the details without by voice distracting from the operation or examination, to devote more to the patient, time and not to the routine, "- says Dmitry Dyrmovsky, CEO of the CST group of companies.

Voice2Med allows you to fill out medical documents in real time, converting the doctor's voice into text. During the examination, the doctor, using a special microphone, dictates information that is instantly decrypted and transferred to the open protocol of the medical information system. The decision, according to the company, saves more than 20% of doctors' time to prepare protocols. The high quality of speech recognition technology that is used in the product (recognition reliability reaches 97-98%) and the ability to connect thematic language models for various medical specializations ensures that at the end of the examination there is a fully completed medical protocol that requires a minimum of corrections. In addition, support for voice commands significantly increases the ergonomics of the workplace.

All real-time data is synchronized with information systems. The product can Voice2Med be installed in the closed circuit of the customer, ensuring the safety of medical and personal data. To simplify the implementation process, boxed solutions for clinics and hospitals have been developed. The composition of both the stationary and mobile workplace of the doctor, in addition to software and training webinars, includes a professional headset: a comprehensive solution eliminates the need to independently select the optimal headsets - everything is available in a single solution. Voice2Med is included in the register of Domestic Software.

Initially, the introduction of a stationary version of the Voice2Med was implemented in Moscow, which became possible thanks to a single digital healthcare platform developed by the capital's social development complex and the Department of Information Technologies. The product is already used by doctors of radiation diagnostics: it works in 80 medical institutions in Moscow. The Center for Diagnostics and Telemedicine recorded in medical organizations of the Moscow Department of Health a reduction in the 22% of time for the preparation of medical protocols: before the introduction of voice input, filling out the protocol took 5-6 minutes, after - 3-4 minutes. Radiologists in Moscow have already filled more than 100 thousand medical protocols with the help of a voice. The decision is also applied in Mordovia, Udmurtia, Murmansk, Kemerovo, Irkutsk, in more than 65 regions of the country, continues to develop and scale.

2021: Ability to fill out medical documents by voice vote

Speech Technology Center Group (part of the Sberbank ecosystem) on August 12, 2021 presented the next product release Voice2Med with the ability to voice fill out medical protocols based on artificial intelligence.

Voice2Med allows you to fill out medical documents in real time, converting the doctor's voice into text. radiologist CT MRI PAT ultrasound diagnostics Two dictionaries of medical terms have been added to the available dictionaries of general practitioner and pediatrician,,, (positron emission tomography), pathologist, physician and endoscopist. This is a dictionary surgeon used when filling out patient diaries, operation protocols, in emergency emergency rest, and a cardiologist's dictionary used during cardiological examinations. To create dictionaries, more than 500 thousand impersonal protocols were processed, which made it possible to ensure the accuracy of term recognition of more than 93%.

Now Voice2Med also supports dictionaries of medical terms for surgeons and cardiologists, which allows you to record details in your voice, without distracting from surgery or examination, to devote more time to the patient, and not to routine paperwork. In addition, to simplify the implementation process, boxed solutions for polyclinics and hospitals have been developed: Voice2Med.Hospital and Voice2Med.Polyclinic. In addition to software and training webinars, the workplace includes a professional headset: a comprehensive solution eliminates the need to independently select optimal headsets - now everything is available in a single solution, "said Dmitry Dyrmovsky, CEO of the CST group of companies.

Voice2Med is already used by radiation diagnostics doctors: it works in 80 medical institutions in Moscow, 35 medical institutions in Mordovia, as well as in Udmurtia, Murmansk, Kemerovo, Volgograd and continues its development. During the study, the doctor, using a special microphone, dictates information that is decrypted and transferred to the open protocol of the medical information system. High quality speech recognition provides a fully completed medical protocol requiring a minimum of adjustments. The Center for Diagnostics and Telemedicine recorded in medical organizations of the Moscow Department of Health a reduction in the 22% of time for the preparation of medical protocols: before the introduction of voice input, filling out the protocol took 5-6 minutes, after - 3-4 minutes.

Continuous work is underway to expand the list of domestic operating systems that are supported by Voice2Med. For August 2021, these are various versions of Astra Linux and Alt.

2015: Description of the work and features of the Voice2Med

The complex for voice filling of medical documents "Voice2Med" uses the technology of recognition of natural combined Russian speech. Voice text input technology allows you to fill out medical documents in real time: during an examination, a doctor using a headset dictates information that is instantly decrypted and automatically transferred to an open protocol. The quality of speech recognition and the ability to connect thematic language models ensures that at the end of the examination procedure, a fully completed medical protocol requiring a minimum of corrections is available.

The system is valuable for ultrasound diagnostics doctors, surgeons, dentists and many other specialists whose hands are occupied during the examination or treatment of patients. "Voice2Med" integrates with medical information systems (MIS), saves doctors time and ensures the reliability of data.

the treatment process and preparing medical documents takes up a significant portion of the time allowed to examine a patient, even in high-automation clinics - more than 30% for a paediatrician and more than 40% for a GP. The first pilot projects have already shown the high efficiency of speech technologies, which allow the doctor to focus on the patient, and not on documents, while also improving the quality of information recorded in electronic medical records, "commented Dmitry Viktorovich, General Director of the CST Dyrmovsky.

Job Description

• During the examination, the doctor dictates the text. The spoken data is recognized in real time and automatically transferred to the protocol opened by the doctor.
• At the end of the procedure, the doctor already has a fully completed protocol, i.e. there is no need to fill it out manually.
• It is not necessary to store information (dimensions and other important parameters of the study) in memory or fix on paper. It is enough to talk and Voice2Med himself will add all the necessary data to the document.

Advantages

• The quality of speech recognition reaches 99.7%
• All medical terms and expressions are recognized
• The system is easily adaptable to the pronunciation of a person and is taught to any topic and word base
• Voice2Med adapts to the speech of a particular doctor, taking into account personal hesitation, possible defects in diction and emphasis
• The system is able to recognize and remove from the protocol parasitic general vocabulary related to conversations of personnel in the room, conversations with the patient, etc.
• Voice2Med supports the use of professional headsets, so that extraneous voices and noises do not get into the system
• The solution easily integrates with third-party medical information systems (MIS), allowing for fast and easy implementation