| Go to: | Welcome | Table of Contents | What's New | Registration | Database |
Technical Items
PRODUCTS
The Company has two main product groups that develop, market and support
products in two distinctive markets.
The Medical Products Group sells, markets, and supports the Kurzweil VoiceMED(R)
and Kurzweil Clinical Reporter(TM) product for the medical market. The
Medical Products Group consists of clinical reporting software systems which
utilize voice input to create structured documents.
The PC Applications Group sells, markets, and supports the Kurzweil Voice for
Windows products to the general personal computer marketplace. These products
allow users to voice-enable Windows-based PCs and software applications, and to
enter data and create text through dictation.
Medical Products Group
The Medical Products Group is dedicated to developing, marketing and supporting
medical reporting software systems for clinicians. These systems are used to
support the direct voice dictation of clinical charts and reports by voice. In
addition, these products can also be implemented as a front-end for a
Computerized Patient Record (CPR), utilizing an interface that extracts clinical
data from the chart to populate the clinical database. The goal of the Medical
Products Group is to support high-quality clinical data management while
expediting the natural workflow of the clinicians providing for patient care.
The Kurzweil VoiceMED and Kurzweil Clinical Reporter products both have up to a
60,000 word vocabulary capacity, and run on Intel-compatible personal computers
under the PC/DOS and Windows operating systems. With systems in use at over 500
medical institutions, the Company believes the Kurzweil Medical Products Group
is a leader in the rapidly emerging market for clinical reporting systems using
voice input technology.
The Company's medical products have three essential elements which the Company
believes make them both useful and cost effective.
The Company believes its Medical Product provides a solution for medical
reporting in six medical specialties:
In March 1996, the Company announced a new Window-based system for creating
complete medical reports by voice. The Kurzweil Clinical Reporter system
integrates medical knowledge bases and a structured report generator and
utilizes the Kurzweil Voice for Windows (release 2.0) voice product as the voice
interface to collect patient data and produce medical reports. The Company
commenced shipment of product to customers in July 1996.
The Clinical Reporter product is available for emergency medicine, radiology,
pathology, and primary care. The Company also introduced Clinical Reporter for
cardiology in March 1997. These products can be run on laptop computers and can
interface mobile monitors ("Cruise Pads"). The product can also support mouse
and pen input devices.
The Clinical Reporter system runs on a Pentium 133 MHz system with 32 megabytes
of RAM on the Windows operating system. The Company currently anticipates a
release of the Clinical Reporter system for the existing knowledge bases on a
system that will run on Windows 95 and NT platforms in the second quarter of
fiscal 1998.
In December 1995, the Company was awarded a $2 million grant by the Commerce
Department's National Institute of Standards and Technology (NIST) Advanced
Technology Program (ATP). The award provides funding for the research and
development of medical record documentation systems
using open systems standards for linking with other healthcare information
systems, and applying user interface technologies including large vocabulary
speech recognition and pen input combined with flexibly structured knowledge
bases designed to facilitate ease of use. Medical record documentation systems
are a prerequisite to the creation of useful clinical databases for outcome
studies, cost-effectiveness studies, and other efforts to improve both the
quality and the cost-effectiveness of medical care in the United States.
The technology developed under the ATP grant is designed to be capable of
accepting clinical information from the care giver (physician exam notes, for
example) and feeding back information, such as clinical treatment guidelines,
from central data repositories.
Product sales to military and veterans hospitals owned by the United States
government totaled $3,525,000, $1,127,000, and $541,000 or 28%, and 12%, and 6%
of the Company's total revenues in fiscal years 1995, 1996, and 1997
respectively.
PC Applications Group
The PC Applications Group sells, markets and supports products used in the
general PC marketplace. The group is focused on becoming the market leader in
voice-enabled PCs and PC applications. The PC Applications Group's key
objectives are:
Kurzweil VOICE for Windows Release 1.0 incorporated a new version of the
Company's large vocabulary, speaker-independent, discrete speech recognition
technology. Users could choose either a 30,000-word or a 60,000-word active
vocabulary. Kurzweil VOICE for Windows Release 1.0 featured an on-line
dictionary including acoustic recognition models and spellings, for a total of
200,000 words. The system is speaker-independent in that most users do not have
to "train" the system on their voice to achieve satisfactory initial accuracy.
The Company announced Kurzweil VOICE for Windows Release 1.0 in April 1994 and
shipments to customers began in July 1994.
Kurzweil VOICE for Windows supports voice input for dictation, which enables the
user to create text and enter data simply by speaking; and navigation, which
controls the Windows operating system and Windows-based applications on a
command and control basis.
Kurzweil VOICE for Windows version 1.2 was released in November 1994 and
shipped with the IBM MWave Windsurfer sound board as its platform. In June 1995,
the Company released Kurzweil VOICE for Windows version 1.5 which included
improved recognition accuracy, continuous digits, increased usability and
shipped with the Spectrum FX sound board.
In April 1996, the Company announced Kurzweil VOICE for Windows Release 2.0
which runs on the industry standard Sound Blaster 16 platform sound board. This
software only release combines a new, more accurate speech recognizing engine,
Windows 95 and network compatibility, and improved speaker independence and
usability.
In November 1996, the Company announced the following product releases:
Kurzweil VOICE Release 2.5 ( Professional Edition ) has a 60,000 word active
vocabulary version, with a suggested retail price of $595. This product is
designed for business people and other professionals with extensive and/or
unique vocabulary requirements. Kurzweil VOICE Release 2.5 (Personal Edition)
has a 30,000 word active vocabulary, with a suggested retail price of $295. Both
versions of Kurweil VOICE Release 2.5 have on-line dictionaries, including
acoustic recognition models, and spelling for a total of 200,000 words.
Kurzweil VOICE Release 2.5 enables PC users to combine voice input with the
keyboard and mouse. The release supports voice input for navigation, which
controls the Windows operating system and Windows-based applications on a
command and control basis, as well as dictation, which enables users to create
text and enter data simply by speaking into a microphone attached to their PCs.
The Company believes Release 2.5 improves user productivity through increased
recognition accuracy and throughput, and through the integration of its
navigation and dictation capabilities into the Windows operating system
including the leading Windows-based applications. The product is
speaker-independent, with an initial "out-of-the-box" accuracy rate which can be
up to 90% or higher. The product also automatically adapts to the user's speech
and language patterns over time, which can boost the ongoing recognition
accuracy rate up to 97% or higher, with easy error correction.
In addition to the price/performance benefits of the Personal and Professional
Editions, Kurzweil VOICE for Windows Release 2.5 includes five enhancements
which ease-of-use:
In June 1996, the Company announced the release of Kurzweil VoicePAD for Windows
1.0 available as a download. This evaluation version of Kurzweil VoicePAD is
available as a download over the Internet as shareware from Kurzweil's site on
the World Wide Web (@ kurzweil.com). Through January 31, 1997 over 45,000 people
have downloaded the shareware version. This version has an active vocabulary of
12,000 commonly spoken words with an additional customized vocabulary capability
of 500 words. The free shareware release is a limited version that restricts
dictation to only 2,500 words. When the word limit is reached the program no
longer operates and the user is encouraged to purchase the Company's product
offerings.
In November 1996, the Company introduced a version of Kurzweil VoicePAD for
Windows Release 1.0. This version of VoicePAD enables the user to seamlessly
integrate voice input with the keyboard and mouse, creating a natural and
intuitive approach to document creation. Like the shareware version, this
product combines the Company's latest discrete speech recognition technology and
allows the user to create text and enter data simply by speaking into the word
processing application. In addition, the user can navigate through the
application on a command and control basis.
Kurzweil VoicePAD for Windows Release 1.0 can be used to create memos, letters,
reports and other documents. Users can format text, navigate through the
application's menus and dialogs, change the application's settings, and preview
and print documents using intuitive voice commands. In addition, by integrating
the continuous digit recognition with the discrete dictation capabilities, users
can quickly and efficiently enter telephone numbers, street addresses, zip
codes, dollar amounts, social security numbers and other numeric data into their
documents.
This version of Kurzweil VoicePAD Pro for Windows Release 1.0 has an active
vocabulary of 17,000 words, with an additional customized vocabulary capability
of 3,000 words. The product does not require any training and the initial
recognition accuracy rate can be up to 90% or higher. Like the other Kurzweil
Voice products, Voice Pad automatically adapts to the user's speech and language
patterns over time, with the capability of improving the accuracy rate up to 97%
or higher, with easy error correction.
DESCRIPTION OF TECHNOLOGY
The Company has developed proprietary technology that incorporates speech
recognition, enabling software and specific knowledge bases.
The Company's approach to speech recognition applies a broad variety of methods,
including techniques derived from statistical, phonetic and linguistic
approaches as well as advanced pattern recognition. The Company's speech
recognition technology is implemented as a collection of software "experts".
These experts focus on different aspects of the recognition process, including
acoustics, statistics, phonetics and linguistics. Results from the experts are
combined to provide a final recognition through the use of the Expert Manager
software module, which coordinates the activities of the other software experts.
If one expert is unable to recognize a spoken word in a particular situation,
the Company's multi-expert approach makes it more likely that another expert
will be able to recognize the word accurately.
The Company's enabling software includes: he Structured Report Generator,
comprised of a database manager, knowledge base shell and word processor; a
knowledge base editor (KBEdit(TM)) program that allows users and developers to
create and edit knowledge bases; and an on-line editor that permits users to
alter the knowledge base while dictating. The Structured Report Generator
includes the ability to respond to a "trigger phrase," which is a single spoken
word or phrase that can trigger an entire pre-defined report segment, with
"fill-in-the-blank" capability for customization. The Company's enabling
software permits users to create their own voice-activated structured reports,
and contains tools for modifying or creating knowledge bases and creating
customized reporting systems.
The Company has developed a set of techniques and development tools which
enables its knowledge engineering organization to develop knowledge bases in the
context of speech recognition. Knowledge engineering is a complex, analytical
process in which a particular field of expertise (a "domain") is organized into
a hierarchical data structure which can be stored and manipulated in a personal
computer. Building a knowledge base for voice reporting requires developing
domain specific vocabularies, appropriate trigger phrases and an underlying
logical framework. Once designed and combined with speech recognition
technology, the knowledge base permits professionals in a personal particular
field to use speech technology to generate reports which reflect their
professional training.
The Company has devoted substantial time and resources in developing and
refining its knowledge bases, and for its Medical Products Group has
collaborated with practicing physicians in relevant fields.
The Company continues its commitment to enhancing and developing the Company's
technology and products. The Company also continues to increase staffing in the
areas of core recognition development, programming development and knowledge
engineering.
RESEARCH AND DEVELOPMENT
The Company has assembled a research and development team that includes
linguists, computer scientists, speech scientists, software engineers, knowledge
engineers and experts in pattern recognition and artificial intelligence
techniques. This team, consisting of 64 full-time employees as of April 15,
1997, is divided into four groups. The Core Research Group, which develops the
underlying speech recognition technology; the Product Development Group, which
develops the user interface, Structured Report Generator, and the application
development tools; the Knowledge Engineering Group, which develops "domain
specific" knowledge bases for the Company's products, and assists applications
developers outside the Company; and the Customer Education Group, which is
responsible for all print documentation, on-line help and tutorials.
The Company is currently focusing its resources on enhancing certain products
that already operate within the Windows operating system; developing a large and
medium vocabulary continuous speech recognition system; and continuing to
enhance its existing product lines.
In fiscal 1996, the Company delivered a voice recognition system operating on
the Unix operating system for GTE as part of a government subcontract. This
CHS.2 contract will voice-enable Unix applications on Sun Solaris workstations
as part of a contract with the U. S. Army extending to 2005.
In November 1993, the Company announced that it had been awarded a grant from
the Advanced Technology Program at the National Institute of Standards (NIST), a
division of the United States Department of Commerce. The grant was a three-year
project which ended on February 28, 1997. As of April 15, 1997, the Company has
received $ 1.7 million in reimbursement under this project. The NIST grant
supported the Company's development of a spoken language interface capable of
controlling personal computer software applications through "natural language"
instruction in combination with a keyboard and a pointing device. The effort is
focused on a new speech recognition interface based on natural language
understanding and continuous speech technology, allowing for the recognition and
interpretation of commands in everyday English from a continuous stream of
words.
In October 1995, the Company announced that it was awarded a second grant from
NIST for a $2 million, two year project. The award provides funding for the
development of medical record documentation systems using open systems standards
for linking with other health care information systems, and applying user
interface technologies including large vocabulary speech-recognition and pen
combined with flexibly structured knowledge bases designed to assure ease of
use. Medical record documentation systems are a prerequisite to the creation of
useful clinical databases for outcome studies, cost effectiveness studies, and
other efforts to improve both the quality and the cost effectiveness of medical
care in the United States.
The technology developed under the second NIST grant will be a two-way system --
accepting clinical information from the care giver (physician exam notes, for
example) and feeding back information such as clinical treatment guidelines from
central data repositories.
The Company believes its VoiceMED and Clinical Reporter products enable
physicians and other healthcare professionals to quickly and efficiently prepare
complete, accurate, immediately available and cost-efficient printed reports by
voice. The products are targeted to specific medical disciplines which utilize
identical and discrete speech recognition technology, and features a customized
knowledge base specific to the medical specialty. With the introduction in July
1996 of the Company's Windows based Clinical Reporter System, the Company made
an effort to reduce the selling and marketing of the DOS-based VoiceMed System.
Kurzweil Clinical Reporter(tm)
Kurzweil VOICE(TM) for Windows
Kurzweil VoicePAD
Go to:
Welcome
Table of Contents
What's New
Registration
Database
May 2, 1997