Think-A-Move, Ltd. (TAM) continues to refine the core functionality of its flagship product, SPEAR®, a domain specific and noise robust automatic speech recognition (ASR) engine. As speech technology continues to evolve at a rapid rate, TAM regularly integrates new features and functionality into the platform in order to address the growing needs of both Commercial and Government users.

The SPEAR solution provides a natural user interface for hands-free, heads-up control, and can be integrated onto phones and tablets running Android or iOS, and PCs with Windows or Linux. The solution can be integrated onto DSPs, is compatible with various Bluetooth accessories, and can run on headsets, smart glasses, wearables, speakers, and other hardware. SPEAR runs locally on the device and does not require a network connection, therefore providing a more secure, portable, and responsive solution.

Current Capabilities include:

  • N-Best Recognition Results for domain-specific post-processing, leading to increased accuracy.
  • Natural Language Understanding to determine intent of the requested command.
  • Text to Speech for providing a spoken response to the user.
  • Knowledge Base Indexing for voice enabled navigation of manuals, flowcharts, and other documents & literature.
  • Alert and Reminder Manager for reducing cognitive load and increasing efficiency.
  • Framework to Monitor External Sensors that support voice inputs.

TAM continues to innovate as shown in the following list of recent enhancements below:

  • RNN-Based Language Data Generation for reduced input training data requirements.
  • Improved DNN Acoustic Model for higher accuracy in noisy environments.
  • Improved Language Model equipped with a Garbage Lexicon for reduced insertion error rate.
  • Improved Voice Activity and End-Point Detection for greater accuracy and faster response time.
  • Grammar Switching between domains without the need to reinitialize SPEAR.
  • Semantic Parsing for mapping tween what is spoken using natural language and what is meant.
  • Grammar-On-The-Fly for supporting user-defined speech recognition task.
  • Keyword Spotting listens for specific words/phrases and triggers user-defined actions.
  • Voice-Enabled Form Filling for hands-free data entry and inventory management.
  • Conversion Calculator for quick access to unit conversion and performing other calculations.
  • Professional Services offered for acoustical environment characterization, audio capture, playback, gain management, speech recording best practices, system specifications, competitive benchmarking, etc.

Comments are closed.