Intro

Imagine you are walking down the street and come face to face with a cat-like animal. How do
you know that it is a cat and not a tiger? In most cases, your senses quickly collect a range of
visual data (body size, ears, how “scary” the eyes look, tail length, paw size, etc.) and make an
overall assessment from this data, which will lead to the conclusion that the animal is indeed a
cat and that there is nothing to fear. Can a computer with AI technology recognize a cat in the
way that humans do? The drastic increase in use of video data (photos, videos, etc.) in this era of big data is leading to
an explosion of interest in (and research on) “computer vision,” or technology related to a
machine’s “vision.” In particular, significant progress being made in computer vision through the
advancement of deep learning technology is leading to an outpouring of tangible and
meaningful research outcomes.

What is
Computer Vision?

There are six major areas in computer vision research: detection (searching for a target/subject
from video data), segmentation (division of video data into several pixel groups), recognition
(identifying a target), classification (separating out based on type), enhancement (increasing
quality [e.g. video resolution]), and restoration (of damaged video data).

One area that is an especially dramatic display of the effects of computer vision technology
(which is in turn armed with big data and AI technologies) is the financial sector. This sector, which collects and processes massive amounts of all kinds of data, is perfect for
applying computer vision technology to the processing of documents and image data to
maximize automation and streamlining. The most common examples are automatic recognition,
classification and analysis of document images, customer data and financial products.

Future Research

HIT is researching and developing diverse computer vision technologies (classification,
recognition, segmentation, detection, etc. of people, objects, letters). A key research area is the
development of OCR (Optical Character Recognition) technologies and services that use these
technologies to recognize and process documents. There is an unlimited number of areas in the financial sector to which OCR technology can be
applied, while automation can maximize convenience and efficiency by saving time and money.
Also, based on the increasing prevalence of contactless activity, facial recognition technology is
growing in importance as a remote means of recognizing customers. As such, efforts are
consistently being made to secure foundational technologies. If the scope of facial recognition is
expanded from (actual) people to currency, services related to automatic recognition and
classification of currency worldwide will advance significantly. HIT is focusing on technology research and service development (via computer vision
technology) of all areas that can revolutionize financial services. Through such efforts, we hope
to become a leader not just within HFG but among agents worldwide in the area of financial
computer vision technology.