Extracted from wikipedia:
“ … During his time as an undergraduate, David Collins worked at the Pennsylvania Railroad and became aware of the need to automatically identify railroad cars. Immediately after receiving his master’s degree from MIT in 1959, he started work at GTE Sylvania and began addressing the problem. He developed a system called KarTrak using blue and yellow reflective stripes attached to the side of the cars, encoding a six-digit company identifier and a four-digit car number. Light reflected off the stripes was fed into one of two photomultipliers, filtered for blue or yellow. …”
This is an explanation about the origin of bar codes.
All the technology, standards and software that was originated around bar codes was about solving a problem which can only be described as: “automatically identify”
Automatic identification is the key concept for interactive systems, in order to improve time optimization, reliability and ergonomy to the manual data entry work.
A good example of automatic identification is fingerprinting. Finger prints are necessary in order to identify human beings. No fingerprint is the same as another and that type of data is much easier to store than DNA samples.
Implementing a system for human identification requires the following steps
1. take a man
2. get a unique characteristic from him: friction ridges of his fingers
3. make a print of this characteristic: fingerprint
4. store this fingerprint and put it together with the man’s data
5. in the future if we need to identify this man, we can cross compare his actual fingerprint to the one in our database.
Fingerprints can also be used as bar codes to identify human beings. Many buildings use this system in order to grant or deny access to a building or specific room, improving time optimization, reliability and ergonomy.
Image fingerprinting technology introduces the previous concepts into an artificial vision field to solve the same kind of problems but related to image identification and recognition.
Fingerprinting for images is calculated using Haar Wavelets. This computes different functions at diverse resolution while storing the results into a vector. This type of technology allows us to compare the person’s fingerprint with an image, for example a picture of the person, with a simple click of a button.
The application fields where we implemented Image fingerprinting are:
1. Identification of Hotel advertisements on a travel magazine to get related offers using an Online Travel Agency system. In this case we could not modify the advertisement by adding a bar code because it was payed by the hotel itself.
2. Identification of paintings for a museum to assist the visitant with related information improving the user experience. In this case we could not include a bar code on the paintings due to obvious reasons.
Both applications were developed for mobile platforms, the first was iOS and the last Android.
In the next video we show an Android device taking a picture of many painting reproductions and showing related info using public websites.