That's pretty inaccurate in your example at 5:59 - there where software counted 317 people. I took screenshot and counted manually (you can do the same) and I can recognize and mark 403 people in the brown area, counting also people whose heads are covered but some other parts are visible (shoulder, for example). I count crowds in Serbia manually for over three decades, and many times I witnessed AI not yet being accurate enough. Naked eye still rules, but of course, good photo or video helps very much.
Hi! The best way to reach out is through LinkedIn. You can add me there (Florian Matusek) or join our group to start a discussion: www.linkedin.com/groups/12487474
Which company video analytics you are using
That's pretty inaccurate in your example at 5:59 - there where software counted 317 people. I took screenshot and counted manually (you can do the same) and I can recognize and mark 403 people in the brown area, counting also people whose heads are covered but some other parts are visible (shoulder, for example). I count crowds in Serbia manually for over three decades, and many times I witnessed AI not yet being accurate enough. Naked eye still rules, but of course, good photo or video helps very much.
What about the personal inside that drive in with crowd into a building
An email please
Hi! The best way to reach out is through LinkedIn. You can add me there (Florian Matusek) or join our group to start a discussion: www.linkedin.com/groups/12487474
Which company video analytics you are using