Person re identification is one of the most challenging field of research, and yet it is very important in computer vision. Due to its challenges it still has many areas which are open for research. It is usually applied to the re identification of a person appearing on one camera as the same person identified previously by another camera on the same camera network. Person re identification is normally applied in security surveillance cameras and has always focused on re-identifying a person or an object from multiple cameras on the same camera system. However this research focused on modifying the person re-identification algorithms to enable counting of re-identified people. This has an application in consumer buyer behaviour by enabling regular customers to be identified and some statistics and analytics of data (tables, graphs, and pie charts) to be developed based on the frequency of the customers' visits to the shop. These statistics can help businesses to know whether they are developing customer loyalty or losing their customers. The significance of this research is the application of counting to person re-identification. Face recognition was used for person re identification since it is more accurate than the re identification of the whole body. The surveillance cameras at the shop’s point of sale can be used to capture the customers’ faces, recognise and identify them through matching with the faces in the database and then determine the frequency of their visit to the shop by counting the number of times they come. Tests were conducted which proved significant accuracy of the counting algorithm. However as future work, face tracking can be combined with face recognition for counting to be more robust and cameras of greater quality for improving the recognition rate