This system will address the current issue of COVID-19 and decrease the spread of disease by detecting the face mask, social distancing, and handshake detection. On spot detection of these measures will help organization to have check and balance on implementation of Standard Operating Procedures (S
CovidDefence
This system will address the current issue of COVID-19 and decrease the spread of disease by detecting the face mask, social distancing, and handshake detection. On spot detection of these measures will help organization to have check and balance on implementation of Standard Operating Procedures (SOPs) at organizations and public places. Our proposed system will address the current issue of COVID-19 and decrease the spread of disease by detecting the face mask, social distancing, and handshake detection. On spot detection of these measures will help organization to have check and balance on implementation of Standard Operating Procedures (SOPs) at organizations and public places.
Objectives:
COVID-19 is an infectious disease that is caused by the Coronavirus associated with severe acute respiratory syndrome. Typically, it is spread from person to person by contaminated surfaces or small droplets. Keeping distance from others and wearing a face mask reduces the risk of transmitting the virus. While a lot of research has been done on face mask detection and social distancing, but there is a room need of improvement in accuracy. Besides this these existing systems lack of detection of handshakes feature as well. Apart from this these systems are tested on cameras rather than CCTV. To address these issues, we have proposed a system “CovidDefence”. Our proposed system will use an annotated video-based dataset “Face Mask Detection Video Dataset” having 4357 video frames, 21941 bounding boxes, 8306 boxes with mask, 13635 boxes without mask, and total of 9GB volume. This dataset will be used for face mask detection and distance calculation. Another dataset including “UOP shakes” will be used for handshake detection. Our proposed method will use transfer learning with YOLO5. Our model will get the images from livestreaming captured by CCTV.
Benefits:
Our system consists of Web applications connected to machine learning models at the back-end through REST APIs. Font end of the application consists of two Tabs. One will display the live streaming coming through CCTV, detect the violated activities and take a snap of the person. The other one Displays the stats about the total violations in the previous 24 hours, week, and month.
| Item Name | Type | No. of Units | Per Unit Cost (in Rs) | Total (in Rs) |
|---|---|---|---|---|
| IGV-IPVD616-5.0MP-W | Equipment | 1 | 27000 | 27000 |
| Gat-6 Cable | Equipment | 1 | 1500 | 1500 |
| HDD 1TB | Equipment | 1 | 9000 | 9000 |
| Power 4A Adabptor | Equipment | 1 | 1500 | 1500 |
| Anxinshi H.265 4k NVR DVR | Equipment | 1 | 14000 | 14000 |
| Tripot | Miscellaneous | 1 | 7000 | 7000 |
| Total in (Rs) | 60000 |
its a simple fitnes app which contain 3 parts where you find the home exercises and...
Smart street lights is a project on intelligent illumination control of street lights to o...
First of all we started with collecting the data for the project by visiting different res...
Description As the title says, our project is about an autonomous del...
Technology is getting enhancements everyday new innovations, ideas and developments are be...