Even though digitalization is increasing rapidly, paper documents, such as books and newspapers, are still widely used. However, the demands to connect these analog and digital worlds are increasing. We already have several methods to do this, such as RFID, barcodes (QR codes, PDF417, and Maxi Code), and watermarking. However, a problem with an RFID is that it cannot be printed on papers. A problem with barcodes is that they occupy space in newspapers or magazines, which may force layout changes. A problem with watermarking is that the data payload is small, and the target media is basically restricted to natural images.
In our new system, information is encoded in a 2D (Two Dimensional) barcode, and is overprinted on the articles of an analog media such as a newspaper or a magazine by using invisible ink. To extract the information, the image with the invisible barcode is photographed using a device such as a webcamera equipped with a blacklight LED (Light Emitting Diode) that makes the invisible ink visible, and the information is extracted using a program in the device. Since the ink is completely invisible in normal lighting, the invisible 2D barcode can be printed anywhere on the printed materials without affecting their appearance, even over text or on colored materials, and paper media in the analog world can be bridged to a digital or virtual world.
This project consists of two major technologies, an extracting technology and a layout technology. In the extracting technology, the image with an invisible barcode is photographed using a device such as a webcamera equipped with a special LED that makes the invisible ink visible, and the information is extracted using image processing technology. In the layout technology, the location and size of the 2D barcode are optimized for printing.
With a normal barcode reader, it is usually impossible to extract the code information in this kind of barcode picture taken when illuminating the page with invisible ink, since the barcode overlaps with the original article. In our extracting technology, we extract the information from invisible 2D barcodes even if the barcodes are overprinted on the articles in media such as newspapers, and even if the article is in color. In detail, as is shown below, we first process the image processing so that the portions with invisible ink are emphasized, allowing us to re-construct the 2D barcode, process it’s error correction code, and extract the information. This extraction algorithm can extract the information in real-time to get the information after a picture is taken by using a PC, and it would take approximately 0.5 seconds even if it were implemented in a cell phone.
Even with our extracting technology, it is sometimes very hard to extract the information because of the quality or characteristics of the invisible ink, paper, and their compatibility. In the layout technology, we select the best locations, sizes, angles, and so on for the printed barcodes within the target areas. We do this by simulating the printing of the invisible barcodes, calculating their response to illumination and the resulting photos, and then simulate the extraction process. It is often possible to achieve a successful extraction rate of almost 100%.