Dictation Tools and Speech Recognition Applications for the Medical Sector


Dictation Tools and Speech Recognition Applications for the Medical Sector
Learn on yet another successful integration of our products

Max Manus AS, based in Norway, is a leading supplier of digital dictation equipment and speech recognition tools for the Scandinavian health sector. For its two leading digital dictation products targeting the health care sector - Dicom 2010™ and Speechmax®, Max Manus AS selected KernelDriver and WinDriver respectively as the driver development tools of choice. The PLX PCI-9052 based central dictation system, Dicom 2010™, enables health care workers to centrally dictate and record large amounts of data, in the hospital setting, whereby the USB-based speech input/output unit, Speechmax® enables local dictation and recording on existing IT platforms. The below case study describes the system make-up and operation, driver requirements and rationality for selecting Jungo's driver development tools for both products.

Both Dicom 2010™ and Speechmax® are dictation tools (ie. record and play back) that aim to make recording in the hospital setting centralized and more efficient. Both products are compatible with Philips' speech recognition applications- Speechmagic®, which can be integrated into these tools to enable automatic transcription. If speech recognition applications are not implemented, transcription can carried out manually subsequent to recording.

The Dicom 2010™, a central dictation system, has been installed by Max Manus in over 70 hospitals throughout Scandinavia. The majority of hospitals in this region aim to digitize their dictation system and move away from the analogue products based on cassette recorders (but restrictions arise due to financial considerations.)

 
.
  Dicom 2010™ in Action

Dicom 2010™ is made up of several components - a dictation microphone, a transcription terminal that communicates via a cabling structure that runs parallel to the LAN and the central sever. All recorded data is saved centrally on the server creating a system that is not dependent on the different client operating systems (making overall management centralized and therefore much easier.) The Dicom 2010™ system provides versatility at the user level, enabling the client to use different units locally - either the USB based Speechmax® enabling local dictation on a laptop (which then interconnects to the Dicom 2010™ server on the network) or as a standalone microphone unit (without a PC, but then which also interconnects to the Dicom 2010™ server on the network ).

Designed to handle large workloads, the system runs as a global service on a Windows server in a LAN/WAN environment. A specially designed expansion card (called MPC-4) residing within the server controls a fibre-optic bus, and provides a ring buffer for a number of simultaneous speech channels. A number of nodes are connected to the bus, each of which supports a maximum of eight sub-nodes. Each sub-node can be either an analogue-to-digital converter or a digital-to-analogue converter, or both. The system is modular, and normally consists of multiple servers communicating over TCP/IP in a LAN/WAN topology. The transcription terminal connected to the via cabling to the central server.

Several client applications are available for communication with the servers. These can run as either a PC client or a terminal service (e.g Citrix).

  .

  Dicom 2010™ System Architecture Diagram

Technical specifications for the system are as follows:
Bus type 1: PCI (32 or 64 bit)
Chip set: PLX 9052 or PLX 9030 depending on version
Bus type 2: MM bus: 2Mbps synchronous fibre-optic
Number of channels: 28 x 64-kbps (simultaneously)
Sound format: CCITT G.711
Operating system: Microsoft Windows NT 4.0/2000/XP (Server version)
Driver development tool: KernelDriver v5.22 (PCI)

The Dicom 2010™ system software is therefore based on different levels. The server provides a global service and utilizes its KernelDriver-based driver to communicate with the PLX PCI-based hardware. The user interface is made up of firmware - the user unit with microphone and transcription terminal at one end and server at the other end. (These units are based on Motorola microcontrollers). There is a unit between the server and "microphone and transcription unit" that controls data to and from the server.   .

  Dicom 2010™ Server Architecture Diagram
.

  The Dicom 2010™ Microphone Hand Set

The Dicom 2010™'s driver development requirements were to interface with PCI with I/O of 128 bytes I/O and Memory of 128 bytes. Max Manus based the system on both the PLX PCI 9052 chipset and their in-house developed expansion card, and due to heavy performance requirements on the system - very wide I/Os, 2Mbps synchronous on the fiber optical side placed heavy reliance for data transfer on the card - resulting in the need for a high driver performance for hardware-to-operating system communication. There was no IRQ (Interrupt ReQuest).

"Using Jungo KernelDriver for the Dicom 2010™ driver development was a natural choice for us due to previous experience with the kernel DDK. Also, as we were porting from ISA to PCI, this was made easier by using a Mini driver," said Morten Schulte Software developer at Max Manus. "Furthermore, developing our own driver would certainly have been much more costly and time-consuming than using Jungo KernelDriver. I state this based on previous in-house driver development experience."

Max Manus' also delivers a solution for PC-based dictation - Speechmax®. Speechmax® is a USB-based complete speech input/output unit. Building on the use of existing IT platforms, the system is ideally suited to smaller hospital/clinic environments. Speechmax® is can be used as a supplement to central hospital dictation systems and other computer systems and is compatible with the Dicom 2010™. The unit consists of a microphone, keyboard, a display, AD/DA converters, memory, microcontroller and USB chip. The software provided with the unit is a standard executable for windows platforms with a windriver-based driver in the kernel that directly accesses the hardware. It operates locally, and is not dependent on the host's resources. Data in both directions is transferred in bulk mode.

  .   The Speechmax® Microphone Hand Set

Technical specifications for the unit are as follows:
Bus type: USB
Chip set: Philips PDIUSBD12
Sound format: CCITT G.711
Operating system: Microsoft Windows 98/ME, 2000/XP
Driver development tool: WinDriver v5.05 (USB)

Max Manus AS is Scandinavia 's leading supplier of digital dictation equipment and speech recognition tools for the health sector. Drawing on 40 years of close collaboration with Philips Speech Processing, the company has gained substantial expertise through in-house development and marketing of advanced speech processing products. Max Manus is also Scandinavia's principal supplier of speech recognition tools. A key to the company's market position is the compatibility of its different products and the ease with which they can be integrated with hospital systems. Max Manus also supplies turnkey dictation solutions. The company's line of analog tools includes desktop models and pocket machines. It also delivers state-of-the-art digital dictation machines.

Max Manus AS can be contacted at: +47 22 30 95 11
Email: aase.nilssen@maxmanus.no
For more information go to: http://www.maxmanus.com

 

  |  Corporate Index  |