AI Solution in the Restaurant Industry


A restaurant has a high amount of walk-ins and generates a major portion of its sales by selling tea. They have a lot of customers who stand outside the store and drink tea.


Since there are a lot of teas sold, they are not able to keep track of how many teas are sold every day. The problem also persists in understanding how much time each customer spends at the restaurant.


To develop a solution which counts the number of teas sold, the number of customers who walked in, and the dwell time of each customer.


We collected videos of the restaurant, focusing on the food counter, and labeled cups of tea for custom object training.

Our algorithm counts the number of cups and the number of sales, while another camera angle counts the number of persons who walked in by assigning them a temporary random ID and measuring the time they spent. This information is updated in real-time on the dashboard.