| Idea Name: |
SSTRUM |
| Slogan: |
The Operator That Doesn’t Sleep |
| Supervisor Name: |
Muhammad Waseem Iqbal |
| Supervisor Designation: |
Assistant Professor |
| Supervisor School: |
Military College of Signals |
| Supervisor Department: |
Department of Information Security |
| Contact number: |
Muhammad Waseem Iqbal |
| Email ID: |
[email protected] |
| Abstract: |
| SSTRUM is a complete machine learning based solution, which automates the process of call tracking in Pashto language by listening to pre-recorded or real-time calls.
|
| What is the unmet need in society that your idea will fulfill ? |
| It is hard for human operators to listen to hours worth of calls manually with concentration, and chances are they might miss some important information during the listen-in. With SSTRUM, we are automating the entire listening process and the operator will only be needed to understand the context. |
| Who needs it ? How many would benefit ? |
| SSTRUM will benefit security agencies, sensitive organizations, and the corporate sector since it provides the facility of upgradation of the dataset according to the need of an organization. This will be beneficial for companies in keeping their communication secure while protecting their privacy. |
| How will the solution works |
| SSTRUM will act as a standalone device configured on Raspberry Pie 4, attached to an LCD. It will contain a specific list of words marked as suspicious and non-suspicious in the dataset. Whenever call audio is fed to the system, our machine learning algorithm will track the audio and detect if it contains any word marked as suspicious. Any audio containing suspicious words will then be treated as suspicious audio for further action taken on it by human operators. |
| Who are your competitors ? How is your solution different |
| No such automated speech recognition system has been developed yet for Pashto language that can be adaptable in terms of datasets, making this system unique in its working. It will be an easier, time-saving, and cheaper method to track speech as compared to the conventional ones. |
| Status: |
new |
| Entry Date & Time: |
2020-12-19 (0933) |
