| Idea Name: |
Key-word Spotting |
| Slogan: |
Finding the presence of words in an audio speech. |
| Supervisor Name: |
Imran Hafeez Abbasi |
| Supervisor Designation: |
Instructor |
| Supervisor School: |
CAE |
| Supervisor Department: |
Avionics |
| Contact number: |
|
| Email ID: |
[email protected] |
| Abstract: |
| Keyword spotting is a speech processing technique that identifies key- words in utterances. Real time solution to automatically detect specific words from speech directly using acoustic-phonetic units |
| What is the unmet need in society that your idea will fulfill ? |
| Word spotting or speech recognition is a difficult task due to the high variance in input signal due to a variety of factors such as gender differences, people’s voices changing over time, health conditions etc. Word spotting techniques that are used makes it possible to spot the required word. |
| Who needs it ? How many would benefit ? |
|
This technique can be used by security/intelligence agencies in identifying the potential threats. It will provide security measures in a way that if any query word is spoken at any technical area, it will be spotted and countermeasures can be taken accordingly. |
| How will the solution works |
| The project would be divided into three phases listed below.
• Feature extraction module
• Deep Neural Network
• Handling module
|
| Who are your competitors ? How is your solution different |
| deep neural network technique will be used in this project. Deep neural Network will be trained to directly predict the keywords or subword units of the keywords. This apprach will not require a decoding algorithm and will reduce runtime computation. |
| Status: |
new |
| Entry Date & Time: |
2021-12-14 (0801) |
