System and method for tracking customer movements in a customer service environment

A system for detecting individuals within image frames and tracking individuals between said image frames, said system comprising:

• processing circuitry, said processing circuitry being configured to receive an image frame;
• a server;
• and
• a database, said database storing datasets, tracklets, and customer sets related to said individuals, each of said tracklets comprising at least one dataset, and each of said customer sets comprising at least one tracklet, and said database also storing metadata related to said individuals,
• wherein said processing circuitry is further configured to process said image frame and to is olate at least one individual's image from said image frame,
• and wherein said processing circuitry is further configured to generate a descriptor for said at least one individual's image,
• and wherein said server receives a specific dataset including said descriptor and compares said specific dataset to a plurality of other datasets, said plurality of other datasets being comprised in a plurality of tracklets and each of said plurality of tracklets being designated as being in an active list, wherein said active list is managed as follows:
• for each tracklet:
• adding said tracklet to said active list at a first appearance of said tracklet;
• assigning a time of last appearance to said tracklet, said time of last appearance corresponding to the last time an image of an individual represented by said tracklet was isolated;
• determining an active list time for said tracklet based on a subtraction of said last time of appearance from a current time;
• comparing said active list time to a limit time;
• when said active list time is larger than said limit time, executing a merging step to merge said tracklet with another tracklet in said active list;
• and
• when said merging step is unsuccessful, removing said tracklet from said active list,
• and wherein said server calculates a similarity score between said dataset and each of said plurality of other datasets, producing a plurality of similarity scores.