Explore GEN family of websites, social media experiences, blogs and more.

Unleashing ideas.
Strengthening startups.

Back to search results
    Date submitted
  • 23-Sep-2016

Pattern based Sound Extraction


We plan to commercialize the audio separation technology as either a pre-filter for popular computer transcription systems for improved transcription accuracy or for hearing air systems for feedback elimination and enhanced hearing aid experienced. The first generation of these products will be in a form of Smartphone app that simply forwards audio to our server that processes, in real-time, and forwards only the desired audio stream(s) to transcribers or back to the app. Subsequent generations and products will include scalable systems for cloud computer as well as ultra-low power standalone IC chip or IP core for mobile systems.


Original YouTube URL: Open

Introduction Video

Additional Questions

Who is your customer?

People with hearing loss: 20-40+% of the population.

What problem does this idea/product solve or what market need does it serve?

For MEPC - Modern speech transcribers like Siri and Cortana have difficult time transcribing speeches with background sounds, therefore users will often relocate themselves into quiet places to use the systems. Our product will separate a mixed recording into multiple clear sounds that can be independently identified, filtered, and correctly transcribed. For MFC - Hearing aids are often not used due to annoying amplified audio feedbacks and unintended amplification of background noise mixed in with desired sounds. Leveraging the same core technology, our product will be able to select and amplify audio signals that are relevant to the hearing aid users.

What attributes will make this idea/product successful? Why do you believe that those features will create success?

Our product will enhance the experience of users of speech recognition system as well as hearing aid users by prefiltering (cleaning/amplifying) a recorded audio through separation of multiple independent sounds that can be selectively forwarded or eliminated.

Explain how you (your team) will execute to make this idea/product successful? What gives you (your team) an advantage over others already in the market or new to this market?

Since 2007, the P.I., Young Cho, has sought to start a company based on a number of his ideas in audio and image processing, sensor network, energy harvesting, and power monitoring. He entered MEPC competition in 2013 on the topic of power monitoring on chip and won fourth place. This effort has led to funding of NSF AIR-TT funding and NSF STTR Phase I funding for the topic. Through these experiences, he has learned more effective ways to narrow down needs of the customers to finding the best product concept to start a successful company. The power monitoring project has been modified over the last 3-4 years based on the customer feedback and has been assigned to be led by his former PhD student, Andrew Goodney, who is, now, one of the CS faculty in USC. Leveraging his entrepreneurship experiences, the P.I. has begun a process of starting another high impact business based on an original audio processing project that he has conceived and investigated back in 2007. Our current demonstration system uses a novel algorithm which is shown to be valid and more effective than any other existing systems. This unique algorithm gives us the technical advantage over other competitors. However, the P.I. understands that there is more to a business than just the technology. Therefore, he has put together a technical team for this project is composed of his research students who have expertise in signal processing, embedded systems development, as well as software development. Furthermore, the team is working with a veteran business advisor, John Sweet, who has 20+ years of experience in business development (13 of which was spent at USC Stevens Institute) but has recently left USC to start an independent company.