news.ycombinator.com• Mar 4, 2019• 1 min read
Books on Object DetectionHi everyone,I am curious about how computers capture an object and classify it correctly. I have seen many research papers and videos where we see green rectangles around people/ bicycles /bus , and written right next to it would be - person (98%), bicycle (95%), bus (90%).What’s even more bamboozling to me is how payment apps work. For ex, in India we have PayTM and PhonePe. Merchants have their QR codes displayed on wall. We open the app, scan the QR code using smartphone app, and pay the merchant. Nine times out of ten, we don’t even have to hold the camera properly and with stability in front of the QR code. Just whizzing the camera by the QR code even helps the app to capture the QR code instantaneously. I am amazed as to how does the camera capture the QR code so fast with so little exposure time.I know this is all object detection but I would love to know all the gory details of how these systems work. The algorithms and frameworks that make this happen.Is it all about pattern c