"Current optical tags, such as barcodes, must be read within a short range and the codes occupy valuable physical space on products," said an MIT document describing the new device. "We present a new low-cost optical design so that the tags can be shrunk to 3mm visible diameter and unmodified ordinary cameras several meters away can be set up to decode the identity plus the relative distance and angle," something the document describes as "like a long distance microscope."
This technique of analyzing the light-reflection has a few potential advantages, such as bokodes are much more difficult to fabricate, meaning it will be much more difficult for thieves to use consumer-grade printers to make fake stickers or to print bogus codes on product containers.
Also, it can trump RFID tags in a few ways: RFID tags "are used to determine the presence of an object within a certain range, but do not reveal its location. They suffer from lack of sufficient directionality and interference with neighboring tags. Although RFID tags can be sensed over distance, they create significant security issues. In the case of passive RFID, the reading distance range may be very limited with the use of a reasonably priced reader. Although active RFID might have greater ranges of operation, it requires an on board power source on each tag. However, RFID still has a long way to go before each user is empowered with a simple RFID reader."
Also, compared with an RFID tag, a bokode could be more easily blocked to prevent someone from reading the tag when it's not supposed to be read.
The bokode's theoretical ability to use low-cost consumer-level cameras may address lowering the cost of infrastructure, as well as making it easier to integrate. At a per-tag price of about $5 for the bokode, it will still need huge volumes to become practical. It also is physically somewhat thicker than a traditional barcode.
An unorthodox element of the bokode approach is that it reveals the data in the coding only when a camera is displaying a blurred, out-of-focus image, which is why the data is not visible to the eye. "Intentionally taking photos with the camera out of focus is somewhat unusual," the document said. "The size of viewable barcode pattern is proportional to the camera aperture size. A relatively large lens aperture is required to see a reasonable part of the Bokode pattern. This explains why the Bokode pattern is effectively ‘invisible’ to the human eye, which has a relatively small pupil size of 2mm to 6mm."
"With the Bokode’s optical setup, the information of the barcode is embedded in the angular and not in the spatial dimension. By throwing the camera out of focus, we capture this angular information in the defocus blur formed on the sensor," the document said. "The pinhole is blurred, but the information encoded in the bokeh is sharp. This also means, for the same sized barcode we can potentially pack 1000 times more bits in both dimension, i.e. million times more data."
This potentially means that the image can not only display a lot more total information, but it can show different information if the camera looks at the image differently.
"The requirements and constraints for the Bokode pattern design are quite different from a traditional 2D barcode. Unlike a traditional barcode, the camera only images a small region of the entire Bokode pattern at a time," the MIT paper said. "This visible Bokode region depends on the distance of the camera from the Bokode and the relative view angle."
"Our Bokode pattern consists of an array of tiled data matrices such that at least one tile is always imaged by the camera within the working angle and distance range. Instead of simply repeating the same information in each Data Matrix, we vary the data bits across the tiles so that the camera obtains view dependent information, something not possible with traditional barcodes."
The report also raised the possibility of using bokodes in large signage. "With Bokodes in street billboards, the human eye will see the billboard information but an out of focus camera will capture the Data Matrix indicating a website link." The full report is available at MIT's site.