UWho program: unique persons counter

One of the organizers of the Makeake make-up conference in Bloomington asked a question that often arises from the organizers of any public events and retail store owners: how many people came to visit us?

There are different ways how to count people. For example, different sensors or a picture from a video camera. Josh Conway has developed a more advanced version of the uWho counter based on OpenCV. She recognizes people's faces, so as not to count one person twice.

When a person gets into the camera lens, the program compares it with all persons saved earlier. If no match is found, the person is added to the database, and the counter increases the value by one.
')
If a match is found, the photo is added to this person’s photo collection to train the system and improve the accuracy of uWho.

Each face has two numbers: the left one corresponds to the identifier of the person in the faceClassifier database, and the right one - the number of photos for that person. All information is stored locally and nothing is transmitted to the Internet. Although, theoretically, it would be possible to automatically determine the name of the person in the photo by searching for similar images on Facebook / Google.

The author says that Facebook already has this technology. Well, this program can be used by anyone for good projects. For example, in the club you can record the time of arrival and departure of people, this is a valuable statistic.

Perhaps someone useful this simple utility. Included with it are a few photos of Josh Conway to configure the classifier, so do not be intimidated by too much.

Of course, the use of such systems should be carried out with the knowledge and consent of the objects of observation, as required by law.