- People Tracking starts with identifying people in frames using AI trained in publicly available datasets for humans.
- Once identified, the particular area of interest is passed on to a deeper level module AI engine that is bifurcate an individual’s traits (e.g. face, gender, hair length, shirt / pant color and type, shoes, specs, etc.) and will fine tune it’s detection architecture’s final layer with this data.
- This knowledge will be later used to track individuals having very similar traits and index them using internal confidence mechanism to rule out the false positives