Well done, and interesting topic. At Mitek, we recently announced that our patented mobile check deposit technology has processed over $1T in deposits. We too leverage computer vision, and this technology is probably the biggest computer vision implementation in banking today, with over 5,600 banks and 80M consumers who use it. Additionally, we are leveraging advance machine learning and computer vision technology to extract information and verify the authenticity of identity documents, thus speeding up new account processes and enabling financial institutions (and others) to say “yes” to more good customers, faster.
One thing (among many) that we have learned is that you must address the variabilities associated with visual inputs from “the wild”. People sometimes take a picture of a check or identity document while riding on a bus, or in harsh sunlight, etc. These are where the hard problems crop up.
Great topic. Thanks for the post.