Project 1

Responses for Project 1

1. The social distance detector was fairly effective. The only errors it faced were with distance recognition and people recognition, as sometimes the violation reading didn’t match the actual distance and not every person was recognized by the object detector.
2. I don’t think this could be very effective while still respecting everyone’s right to privacy. In order to accurately track transmission of the virus, we’d need to know a lot, such as whether or not they were wearing a mask, who they’ve already been in contact with (since social distancing doesn’t really apply to those you live with), whether or not they’ve done something substantial to transmit the virus (breathing in another’s direction, touching another, coughing, sneezing, etc.), and a number of other factors. While probably possible, knowing these factors about each individual seems both a difficult venture and a violation of privacy on some level.
3. You could potentially use infrared thermography to capture breathing (it’d be incredibly difficult, but it’s possible). You could also use object detection to see who’s wearing a mask and who isn’t. Additionally, while social distancing is certainly a great practice for preventing transmission, distance itself isn’t necessarily the cause of transmission, so an improved version of the model would keep track of these more direct causes, perhaps in replacement or addition to the social distancing detection.