Speech Recognition Solutions
Professional Custom Speech Solutions
We offer professional services for those clients who expect best-in-class performance. If your business is planning a speech strategy, consider how MatrixHCI can save you time and money while providing quality assurance. Our precision team can solve your speech needs related to audio conditioning, grammar engineering, speech tuning, dialog authoring, and personal assistant development.
MatrixHCI delivers speech-driven solutions across many vertical markets, including services for mobile voice recognition, call center ASR accuracy, medical speech recognition, voice controls for aviation, and conversational speech systems.
Some Sample Speech Solutions
Voice Interfaces for Aviation
Speech is a natural interface that can improve situational awareness and decrease mental workload in a hands-busy, eyes-busy environment.
In aviation this hands-busy, eyes-busy environment is found in the cockpit, the Tower, the Ground Support vehicle, or the UAV (Unmanned Aerial Vehicle) control workstation.
Despite these benefits, speech solutions for the aeronautics domain present tough challenges:
- noisy environments
- specialized vocabularies
- low bandwidths
- bursts of high-rate speech
By providing focused consulting, MatrixHCI surpasses the performance of commercially available speech software.
This achievement begins with the analysis and development of custom acoustic models that are matched to the target environment. Custom grammars are trained to the unique speech features like the vocabularies used in aviation (e.g., delta bravo) and the rapid speech rates to further reduce errors.
MatrixHCI takes grammar engineering to a new level. Ordinary systems use static grammar files. MatrixHCI develops dynamic grammars. Besides being more processor efficient, the outcome is increased recognition accuracy, control, and system scalability. Then we enhance them with an AI module that tracks operational state and vets recognition results against it.
The result is a solution that doesn’t just tell you what they said, it also warns you when that’s not what they meant.
Mobile and Embedded ASR
MatrixHCI understands the constraints of mobile and embedded speech recognition.
Too many of today’s speech solutions start from the premise that you will tether yourself to the Internet. For many applications this isn’t feasible, for others, there are huge security issues with sharing personal data with the Cloud.
MatrixHCI has the skills to analyze and optimize your real-time mobile speech products.
Our services include tuning all aspects of the speech engine for small devices, including cell phones and tablets. We can develop the specialized speech technologies you require, including audio conditioning algorithms for mobile environments, and more effective and efficient grammar strategies for low powered hardware.
Hotel Call Centers
Corporate IVR centers must balance two competing issues: fast transcription of customer calls to support real-time business data mining and purging sensitive customer information to support compliance with personal privacy and security legislation.
MatrixHCI can customize a speech system that meets both goals.
At the heart of the call center solutions are custom acoustic models developed using actual calls recorded onsite. Next, a speech AI (artificial intelligence) module is trained to spot and scrub credit card numbers and other personal information. Last, segmentation testing and tuning are used to convert the audio to searchable text. This solution greatly improves a brand’s ability to monitor and respond to inbound customer issues while safeguarding sensitive data.
Personal Assistants
A little bit of background knowledge can go a long way in simplifying a speech interface.
MatrixHCI can deliver a system that maximizes the customer experience by linking your voice-driven front end to industry standard databases and adding custom AI modules that encode your business logic. With this intelligence, users don’t have to answer the same questions over and over. We can develop dynamic grammars that allow experienced users to short-cut the standard menus and go directly to the services they need. The result is a solution that provides the feel of a personal assistant enhancing your brand’s value.
Speech Solutions for Hosted Voice Portals
MatrixHCI is experienced with scalable “Cloud-based” voice-as-a-service applications.
While the architecture for general software-as-a-service (SaaS) systems is well understood, the knowledge to quickly implement and support a robust speech service is unfortunately not such a commodity.
MatrixHCI provides a fast and reliable way to ramp up a professional speech team. We support clients through the entire life cycle process and guide them in best practices for voice user interface (VUI) design. Starting with the needs and requirements analysis, we identify the key attributes for a successful hosted voice service. We account for audio pre-filtering given the random input sources connecting to the remote application. Once at the recognition level, we offer a proprietary engine or we can work with other commercial speech technologies. And, finally, as your team member, we can handle continued support and benchmarking so your organization can focus on driving product growth.
Beyond the box …
MatrixHCI was created to help customers lower costs and time outlays in pursuit of more robust voice recognition applications.
MatrixHCI is a speech development lab providing complete ASR solutions or speech project consulting in one or all of the service areas below:
• Audio Conditioning
• Pre / Post Filtering
• Audio Segmentation
• Custom Acoustic Models
• Custom Grammar Development
• Code Development
• Engine Optimization
• Advanced Techniques, AI, and Adaptive Routines
• Algorithm Development
• Embedded or Hardware Related Challenges
If you’re not satisfied with your current system’s accuracy, or can’t get the attention you require to optimize your voice interface, please contact our sales department to discuss your voice-to-text needs. We have speech solutions for the most challenging environments.
