Idea Name: |
Imposter Detector |
Slogan: |
You can run can't hide |
Supervisor Name: |
Dr. Ahmad Salman |
Supervisor Designation: |
Assistant Professor |
Supervisor School: |
SEECS |
Supervisor Department: |
Department of Electrical Engineering |
Contact number: |
+925190852559 |
Email ID: |
[email protected] |
Abstract: |
"Imposter Detector" is an AI based innovative solution for surveillance and human detection in areas that are out of bound for public, be it be border areas or out-of-bound industrial areas. |
What is the unmet need in society that your idea will fulfill ? |
Border areas are spread over miles and due to workforce constraints, limitations of cameras on ground, soldiers cannot do effective surveillance and detect human presence causing illegal activities in the border areas. Similarly, industries face difficulties in human detection in no-go areas. |
Who needs it ? How many would benefit ? |
It is needed by military for intelligence gathering and prevention of crimes. Industries need it for human detection in out-of-bound areas ensuring protection of a process and people working there. It will ease the burden on military personnel and will ensure public safety making them beneficiaries. |
How will the solution works |
We will be using yolov3, which is a deep-learning based architecture. We will train the model and then deploy if for human detection. The model will be deployed on Nvidia TX2 board on a quadcopter mounted with on-board camera, which will get the video feed and the deployed model will detect humans. The feed will be transmitted to the ground control station. For indoor areas, CCTV cameras can be deployed with the trained model. |
Who are your competitors ? How is your solution different |
There are no current competitors in the private sector. |
Status: |
new |
Entry Date & Time: |
2020-12-20 (1822) |