Plano, TX, United States of America

Yu-Hung Kao


Average Co-Inventor Count = 1.3

ph-index = 6

Forward Citations = 153(Granted Patents)


Location History:

  • Richardson, TX (US) (1998)
  • Plano, TX (US) (2001 - 2006)

Company Filing History:


Years Active: 1998-2006

Loading Chart...
9 patents (USPTO):Explore Patents

Title: Yu-Hung Kao: Innovator in Speech Recognition Technology

Introduction

Yu-Hung Kao is a prominent inventor based in Plano, TX, known for his significant contributions to the field of speech recognition technology. With a total of 9 patents to his name, Kao has made remarkable advancements that enhance the accuracy and efficiency of speech recognition systems.

Latest Patents

One of his latest patents focuses on implementing a high accuracy continuous speech recognizer on a fixed-point processor. This innovation describes a small vocabulary speech recognizer suitable for implementation on a 16-bit fixed-point DSP. The process begins with sampling input speech at an analog-to-digital converter, followed by applying digital samples to Mel-scaled cepstrum coefficients (MFCC) front end processing. To ensure robustness against background noise, a parallel model combination (PMC) is integrated. The MFCC and Gaussian mean vectors are utilized in the PMC, which provides speech features extracted in noisy environments. This technology modifies hidden Markov models (HMMs) to improve recognition accuracy. The method of computing MFCC involves several steps, including dynamic Q-point computation, Hamming Window application, and complex FFT to power spectrum conversion. Additionally, a polynomial fit is employed to compute log2 in the log filter bank step. The PMC computation method includes steps for computing noise MFCC profiles and performing model combinations in the linear filter bank domain.

Another notable patent is a compact text-to-phone pronunciation dictionary. This innovation achieves a five times compression of a typical English pronunciation dictionary, which usually occupies up to 1,826,302 bytes in ASCII. The compression is accomplished through prefix delta encoding of the word and error encoding of the pronunciation, while maintaining computability.

Career Highlights

Yu-Hung Kao is currently employed at Texas Instruments Corporation, where he continues to develop cutting-edge technologies in speech recognition. His work has significantly impacted the industry, leading to more efficient and effective communication systems.

Collaborations

Kao has collaborated with notable colleagues, including Kazuhiro Kondo and Yifan Gong, contributing to various projects that push the boundaries of speech recognition technology.

Conclusion

Yu-Hung Kao's innovative work in speech recognition technology has established him as a key figure in the field. His patents reflect a commitment to enhancing communication systems, making significant strides in the industry.

This text is generated by artificial intelligence and may not be accurate.
Please report any incorrect information to support@idiyas.com
Loading…