PRE2020 Group 1 numbers: Difference between revisions
Jump to navigation
Jump to search
No edit summary |
No edit summary |
||
| Line 99: | Line 99: | ||
|- | |- | ||
|<!--Col1--> | |<!--Col1-->Expected false positive rate in measurement station | ||
|<!--Col2--> | |<!--Col2-->7% | ||
|<!--Col3--> | |<!--Col3-->Study of similar cough classification algorithm achieved 96.6% accuracy. In this specific case this becomes higher than only the cough test since only 1 test needs to be positive to refuse access | ||
|<!--Col4--> | |<!--Col4--><ref name = "CoronaCoughApp"> [https://doi.org/10.1016/j.imu.2020.100378] Imran, A., Posokhova, I., Qureshi, H. N., Masood, U., Riaz, S., Ali, K., John, C. N., Hussain, I., & Nabeel, M. (2020). AI4COVID-19: AI enabled preliminary diagnosis for COVID-19 from cough samples via an app. Informatics in Medicine Unlocked, 20, 100378.</ref> | ||
|- | |- | ||
Revision as of 09:00, 15 October 2020
- Range microphone: 10x10 meter. [5]
- Height microphone: 2.5 meter since this is the minimum height of the chosen microphone.
- Distance between the detection device and the measurement station: preferably 2 meter but can depend on the available space and which entrance.
- Percentage of infected people: 0.1%, this is the amount of people currently tested positive for corona. Some of these people are in the hospital and do not count, also the people tested positive will be a lot careful coming into contact with others. However there are also people that have corona that are not tested, This will even out and so the same Percentage of infected people was taken. [6]
- Chance of detecting a cough: 87% Korean study
- Time at the measurement station per person: 4 minutes, 1 minute testing, 2 minutes disinfecting, 1 spare minute.
- False positive in the measurement station: 7% becomes higher than only the cough test since only 1 test needs to be positive to refuse access [7]
- False negative in the measurement station: 7% becomes lower than only the cough test since only 1 test needs to be positive to refuse access.[8]
- Random people picked out: after the device detects 2 people 1 random person is picked out, only if there is no row.
- Chance a healthy person coughs close to the camera: 0.08% based on average cough time and waling speed [9]
- Cost of the product:
- €5500 for thermal camera
- €1000 for microphones
- €100 for microphone in cabin
- €400 for camera on pole
- €750 for camera pole
- €3000 for booth including computer, software, screens, and interfaces
- €200 for UV light
- Salary of a Medical personnel: €25 and hour gotten from a salary administrator
- Dutch reproduction rate of Corona: 3 [10]
- Number of people entering the Heuvel Galerie during 30 min:
Quiet hours
- Quiet entrance: 24
- Busy entrance: 61
Medium hours
- Quiet entrance: 117
- Busy entrance: 194
Busy hours
- Quiet entrance:233
- Busy entrance:756
| Name | Value | Explanation | Source |
| Microphone array range | 10x10 meter | Range provided by manufacturer | [1] |
| Microphone mounting height | 2.5 meters | Minimum mounting height provided by manufacturer | [1] |
| Distance between detection device and measurement station | 2 meters | Preferred value | |
| Percentage of infected people | 0.1% | This is the amount of people currently tested positive for COVID-19. Some of these people are in the hospital and do not count, also the people tested positive will be a lot more careful coming into contact with others. However there are also people that have COVID-19 that are not tested. This will even out and so the same Percentage of infected people was taken. | [2] |
| Chance of detecting a cough with camera and microphone | 87% | Chance obtained by Korean researchers using a similar system | [3] |
| Time spent at measurement station per person | 4 minutes | 1 minute testing + 2 minutes disinfecting + 1 spare minute | |
| Expected false positive rate in measurement station | 7% | Study of similar cough classification algorithm achieved 96.6% accuracy. In this specific case this becomes higher than only the cough test since only 1 test needs to be positive to refuse access | [4] |
|
|
- ↑ 1.0 1.1 Sorama. Sorama Smart City Listener. [1]
- ↑ RIVM. Actuele informatie over het nieuwe coronavirus. [2]
- ↑ Center for Noise and Vibration Control KAIST (2020). Deep Learning-Based Cough Recognition Model Helps Detect the Location of Coughing Sounds in Real Time. [3]
- ↑ [4] Imran, A., Posokhova, I., Qureshi, H. N., Masood, U., Riaz, S., Ali, K., John, C. N., Hussain, I., & Nabeel, M. (2020). AI4COVID-19: AI enabled preliminary diagnosis for COVID-19 from cough samples via an app. Informatics in Medicine Unlocked, 20, 100378.