Check transactions represent a major information-rich stream of customer data that is available to banks today. However, this valuable source of data is largely untapped due to the difficulties associated with extracting data from check images. This paper will show how it is possible to exploit image processing and pattern recognition techniques to help extract and analyze the information content of check transactions. It will also explore how this information could be used within the context of a data warehouse to provide banks with a better understanding of their customers, allowing them to approach those customers with customized offers for products or financial services. In addition, a case study of a prototype of a check image data mining system is presented demonstrating the feasibility of this technology. The paper will also touch on the implications of this business application to consumer data privacy.
Read full abstract