成員：阮柏愷 郝嘉誠 邱易檠 陳克齊

Origin and overview
At present, home detection instruments aimed at COVID 19 mainly came from the SHERLOCK test (STOP), led by Professor Zhang Feng of MIT, and the detectors of Professor Jennifer Doudna of the University of California at Berkeley. Both of the instruments are using CRISPR technology for detection.
Although this technology can facilitate the citizens to self-detect and reduce the consumption of medical resources in the hospital, it may also reduce the patient data and the information of potential outbreak areas that doctors and epidemic prevention personnel can control. Therefore, our team is committed to designing a platform that can accompany SHERLOCK. After the patients do the test themselves and upload the photos of the test strips to our platform.

Our goals include:
1. Creating a platform to let users upload the result of test strips in the form of photos.
2. Doing data analysis and connect the photo with the basic information of the users who is judged to have COVID-19 disease by our software
3. Combining users' location information of the map to present the potential hot spot of COVID-19. This function help residents living in the area to be more alert.
4. Sending information to the Centers for Disease Control and Prevention through blockchain.

Technical level
1. Test strip image recognition
Because the result of the test strip is relatively simple, neural network and other models can be combined to establish an identification system.
2. Data transfer and database establishment
To create the system, the establishment of a database is necessary. Consequently, the obtained information can be encrypted and stored in the remote server. At the same time, special fields can also be added, such as: location data, confidence rate, etc., to provide more diverse information.

3. Combination with the map of the epidemic
By taking out the location data in the database and cooperating with the Google Maps API, we can create a simple epidemic prevention map for the public, letting them refer to the number of infected people in each area and allow the government to change the epidemic prevention strategy for the area in real time.
4. User interface design
Combine the above functions and create an APP for the public to use. simultaneously, the promotion of APP would be further promoted through agency announcements and policy formulation.

Possible obstacles for this technology
1. Privacy issues
The test results and the location data are relatively personal, so many people would rather to keep their profile unknown to the public. Moreover, if other people have access to these data, it can lead to malicious uses
2. Accuracy
Even if the accuracy of the model reaches 90%, it cannot be said that this method is effective. Further exploration indicators such as "Recall" or "Precision" are needed to determine whether the model is good or bad. On the other hand, the wrong prediction (false positive) may cause people to panic. To resolve this problem, more medical resources will be needed.