recoAnywhere: a vocabulary for all medical disciplines

In the introduction of a voice recognition system on the entire hospital structure at least 6 criteria should be considered to make a right choice: distributed architecture, integration, possibility of use on different clients (eg mobile), correction flow, availability of vocabularies , scalability.

With volumes of enormous reports and the habit of dictating, the imaging departments (radiology, neuroradiology and cervical medicine) were the first to benefit from the advantages obtained by the introduction of voice recognition. Other departments, having verified the results of the application of speech recognition in radiology, now set the same goals. Several institutions are thinking of extending the speech recognition paradigm to the entire hospital structure.

The recomed vocabulary, always updated, includes the terms for the following disciplines: Radiology, Neuroradiology, Nuclear Medicine, Pathological Anatomy, Radiotherapy, Cardiology, Urology, Otorhinolaryngology, Gastroenterology, Gynecology, Neurology, Oncology, Intensive Care, etc.

Contact us for an updated list, or to verify that your discipline is covered.


A homogeneous information system facilitates cooperation between departments and the management of the entire infrastructure. The voice recognition system must be integrated in the various management systems (RIS, HIS, PACS, EPR, etc ...) to optimize existing workflows without changing work habits.


A distributed network solution must allow centralized management of the entire system, ensuring maximum flexibility for administrators. Management and configuration tools must be accessible from any workstation. Every doctor must be able to access their voice profile from anywhere to report from any workstation.


A speech recognition system is a long-term investment and therefore must be able to grow within the organization. It must be scalable in terms of size, vocabularies and even functionality.
It must work on any device, with any OS, on wide area networks (even for thousands of users) on physical or virtual servers, with load balancing and fault tolerance mechanisms.

One Topic, Shared

The accuracy of speech recognition depends largely on the vocabulary and language model. A good system must provide maximum robustness in generic environments and be able to be optimized on individual disciplines.
The multidisciplinary medical vocabulary issued for Italy currently covers more than 90% of the medical terminology used in the entire hospital structure and can be further extended with onsite customizations and pre-adaptations. In the recomed architecture the vocabularies are shared so that the automatic adaptation of each of the doctors can contribute to the performance increase for all.

Multimodal and Multichannel

Each department has its own needs. The use of mobile systems (notebooks, tablets, etc.) is also on the rise in the health sector. A modern voice recognition system must be able to be used on all the necessary devices, in all modes (direct, streaming or delayed), regardless of the client's operating system, guaranteeing all the features and the best performance in any context. It must also allow its use with any acquisition device, from traditional microphones with control buttons to headsets wired or wireless, for those who need to keep their hands free.

Easy Correction

Among the advantages of adopting a voice recognition system, the possibility of optimizing the workflow stands out. It must be accompanied by adequate verification and correction tools, so as to be able to dictate, correct and release the document in a short time and with the highest quality.
© 2024 Reconice. All Rights Reserved. Powered by reconice