The facial recognition can be improved automatically by hooking in mobile phone cell tower tracking (cell tower traffic management) in realtime which I believe the police here in the UK also have, then over time the realtime cctv's can also adjust for differences in the camera sensors, like colour hues and white balance, because not everyone is going to follow the same route at the same speed at the same time as everyone else. Its not hard to do if you join up the technology that is also in use, including cash machine usage, shopping payment systems, car park ticket machines, etc etc. Individually, facial recognition on its own has more weaknesses than if its joined up with other tech where users leave a digital footprint.
If you take the new electricity smart meters being rolled out in the UK and elsewhere, over time, it would be possible to work out the devices in use by the amount of electricity that is used and the pattern or profile seen of electrical consumption.
EG, if you saw a small increase in use of electricity at 4am for a about a minute, you could surmise that someone has turned a light on and possibly gone to the toilet. Over time, you will the same or similar amount of electricity used which could then be used to monitor the toilet habits of the individuals in the property. This could be a measure over the years for prostrate problems if its known a male lives in the property, as an example of how the data could be used to hilight medical problems. If you take a washing machine, this might be hot fill or cold fill, but you could watch the consistent increase & decrease of electricity as the washing machine goes through its cycle, finishing up with a spin cycle. Machines are excellent for consistent demand of electricity. In time, you could even work out what cycle or programme has been selected on the washing machine.
When combined with other devices with their own electrical demand patterns, you could work out everything electrical in use, and also identify or predict faulting electrical devices. So the electrical companies could predict wasteful users as well as conservative users if other factors are known like the type of properties, useful or made easy on new build housing estates.
This in turn can also be combined in realtime with police systems taking you closer to the fictional minority report, so police could be on hand to prevent a situation in the home or in the street, which I would hope they would be there to do, rather than just arrive after an event and pick up the pieces as this is also traumatic for them seeing as they are also human beings.
