Theft Detection Algorithm for Retail

I'm looking for a machine learning expert who can create an algorithm for me that detects theft in real time. This algorithm should be able to identify when a shopper is taking items from both shelves and baskets, placing them into a bag or purse, and subsequently leaving the store without paying.

Key Requirements:
- The algorithm needs to process a live camera feed, so it must be capable of real-time analysis.
- The alert system should trigger only when an item is taken and not scanned at checkout.

Ideal Skills:
- Proficiency in machine learning and algorithm development.
- Experience with video processing and real-time systems.
- Understanding of retail theft patterns and prevention techniques.

Please note: I have the rest of the system implemented; I just need you to develop this specific algorithm. Scammers need not apply.