SECURITY MEASURE SYSTEM FOR VISUAL MACHINE-READABLE CODES
The present specification relates to security measure system for visual machine-readable codes, particularly barcodes and the like Discount vouchers are often provided by stores, such as supermarkets, as a way of inducing selected customers to purchase an item at a discount. The voucher often incudes a barcode that, when scanned in the usual manner, is recognised as reducing the cost of another item that has been brought, or some other reward. These vouchers are often provided on a limited basis, intended to be used on a single occasion by particularly selected customers.
Traditionally, these vouchers have been provided as a printed hard copy, but more recently they may be distributed to a customer electronically, so that they can display them on the mobile phone at the store checkout.
This system is open to abuse however. Some recipients of the voucher upload an image of the barcode to social media or share the image with friends and acquaintances, so that people who haven't been selected to benefit from the voucher can use them (by displaying the same image on their mobile phone at the checkout), and even use them more than once. This can cause an unexpected loss to the store if it cannot distinguish properly-issued vouchers from fraudulently duplicated vouchers.
The object of the present invention is to make duplication of such barcodes more difficult, while allowing properly-issued vouchers to be conveniently read.
According to the present invention, there is provided a security system according to claim 1.
The invention will now be described, by way of example, with reference to the drawings, of which; Figure 1 is a schematic representation of discrete images making up a moving image according to one embodiment of the invention, Figure 2 is a diagrammatic representation of a security system in use, and Figures 3 to 5 are diagrammatic representation of generation and distribution of the moving image, Referring to figure 1, a changing or moving image 25 for use in the security system comprises a number or collection of different individual images 10, 11, 12, 13, 14, which are displayed sequentially as a moving image. Each comprises or includes a barcode 20, 21, 22, 23, 24. One barcode, for example barcode 22, is a genuine barcode that will be recognised by a store scanning system and result in some action being taken by the store scanning system, such as applying a money off voucher.
The other barcodes 20, 21, 23, 24 are pseudo-readable images, specifically in this example fake barcodes, chosen to not be recognised by the system. A typical barcode comprises parallel bars arranging in a specific format, representing a number string, and includes a check digit by which means the scanner system can verify that the barcode is correctly composed.
The fake barcodes may be composed of parallel spaced black bars (of varying widths as in a genuine barcode) on a white background, but are not detected or considered by the scanning system.
The pseudo-readable image should ideally be unrecognised and not considered as a candidate image for any further analysis at an initial or early stage of the scanner detection algorithm, so that the scanner does not allocate processing resources (or only allocates minimal resources) to attempting to decode the pseudo-readable image, and does not delay in reading the genuine barcode when it appears in the sequence. The composition of the pseudo-readable code could be tailored to the particular scanning systems that it is intended to be used with, in order to generate pseudo-readable codes that are unrecognised or rejected with no or minimal processing by the scanning system.
Scanning systems use different strategies to detect candidate barcodes and then attempt to decode them. For example, in a known scanning system, the reader continually checks the view in front of the reader for images that either match one or more designated templates (for example, using gradient map or feature matching template) for possible candidate barcodes, and on encountering a candidate barcode, then attempts to decode the barcode.
For example, the scanning system could initially check for the presence of 'guard bars' at spaced at specified intervals which are present in a barcode that conforms to the UPC standard.
Therefore pseudo-readable image is ideally chosen so that while it looks like a barcode to the human eye, it lacks the characteristics which cause it to qualify as a candidate barcode for further processing. There are many ways that a pseudo-readable image can be varied as compared to a genuine barcode, for example by ensuring the pseudo-readable image does not match the template used by the scanning system.. For example, a fake barcode can be composed of parallel black lines of varying width, as in a genuine barcode, but lacking the specific widths and distribution of a genuine barcode, or lacking a feature such as the 'guard bars' of a genuine barcode.
Less ideally, the pseudo-readable image could be composed to include some characteristic of signifier which means that the scanner system will not recognise the barcode, or will identify that it is not a genuine barcode, such as including a deliberately incorrect check digit.
Referring to figure 2, in order to use an electronic voucher that has been distributed to a customer 100, the customer 100 displays the moving image 25 on their mobile device 102 such as a mobile phone for it to be scanned by a scanner 101 and entered into store system at the point of sale device 104.
As the images 10, 11, 12, 13, 14 are presented sequentially, the scanner attempts to read each image. The fake barcodes 20, 21, 23, 24 are not read by the store system, while the genuine barcode 22 is read, converted to a corresponding code (usually a numeric string) and acted upon to actuate the purpose of the barcode, such as applying a discount on other items the customer 100 has brought.
As previously mentioned, when someone duplicates an existing voucher, this may be done simply by taking an image of the static barcode, displaying it on another mobile device and presenting it at a store for scanning. The moving image makes this much more difficult, as when taking a static image, the person attempting to duplicate it is more likely to take an image of one of the fake barcodes that are presented, rather than the genuine barcode 22.
This also means that distributing the genuine barcode 22, for example via social media, is very much more difficult.
The number of fake barcodes included in the moving image 25 can be increased if necessary, for example twenty fake barcodes could be used with a single genuine barcode. The genuine barcode could be inserted into the sequence of fake barcodes at any point, and its position in the sequence could be randomised for each voucher used in the system, or each time the moving image is generated. The genuine barcode could be inserted more than once in the sequence and some or all of the fake barcodes could also be repeatedly displayed when the moving image is displayed, and their order could also be randomised.
The whole sequence of the moving image could take ideally in the range of 2 to 15 second, typically 5 seconds. Each image could be shown ideally in the range of 0.05 to 0.3 seconds, ideally for about 0.1 seconds. The difficulty of duplicating the image by someone taking a still photo increases with the number of images used (which increases the likelihood of selecting the incorrect image), and as the duration of each image decreases. A range of S to 40 images works well, typically 20 images would be used.
Repeating the display of the genuine barcode increases the chance that the scanner system will be able to correctly read the genuine barcode. The whole sequence of the moving image could display the genuine barcode two or three times. The generation of the moving image could include rules to ensure that the genuine barcode is not displayed twice without a fake barcode between each display instance of the genuine barcode.
The number and duration of each distinct image of fake and genuine barcodes is chosen with regard to the temporal resolution and capabilities of the scanning and decoding apparatus.
Typically, the duration each individual image is displayed for is constant, but different durations could also be used.
The specific method by which the fake barcodes are not read is that they are generated to have incorrect check digits. However, the fake barcodes could be generated to be unreadable by other methods for example because the bars are incorrectly distributed or have incorrect relative widths which do not correspond to numbers and aren't recognised by the scanner system. Less ideally, the fake barcodes could be correctly read, but rejected as being recognised fake barcodes by the scanning system.
Referring to figure 3, a database system 30 stores genuine barcodes to be used in vouchers, and a collection of fake barcodes. The genuine barcodes and fake barcodes will of course be stored separately or have distinguishing data associated with them. An operator 31 assembles the moving image 25 by selecting a genuine barcode and a number of genuine fake barcodes from the database system to create a Graphics Interchange Format (gif) that displays the genuine barcode and fake barcodes in the manner previously described.
This gif is then transferred to a server and on to the customer's mobile device on demand. For example, the voucher may be transmitted to the customer by supplying them with a link to a webpage, which the customer can displays the gif in their mobile device's browser when in store.
Referring to figure 4, the process may be automated by using a rules engine 36 which determines how the genuine and fake barcodes should alternate in display. A server 33 receives all barcodes and the rules engine 36. When barcode is requested, the server 33 assembles all together at run time to create an animation in real time, which is displayed on the customer's mobile device 102 when in store.
The rules engine could also include a method for generating fake barcodes or pseudo-readable codes. There are many ways that the suitable pseudocodes could be generated, for example genuine barcodes could be randomly altered so that known areas of a genuine barcode are changed in a known or random way to create codes that would be unrecognised. The rules engine could include a checking step where detection algorithms are applied to the pseudocodes to check that they aren't recognised as candidate barcodes or generate spurious decoded data by the scanning system.
The genuine product code could be transmitted to the rules engine for dynamic generation of the genuine barcode either on server or (in the example described below) by generation by a secure app on the user's mobile device.
Referring to figure 5, another method of automatically generating a moving image in real time is shown where the genuine and fake barcodes database 30 is stored on a server 34 with a rules engine 37, which determines how the moving image should alternate in display. The server then transmits the genuine and fake barcodes to the user device on demand, and the mobile device 102 uses the rules engine to create the animation at run time.
It will be realised that other architectures of generating and distributing the moving image can be used. The moving image need not be a gif file, but could be stored as a moving image in any format. As an alternative to a database of genuine and fake barcodes, genuine and fake barcodes could be dynamically generated on demand, on a server to be supplied to users, or even on the users' phone with a suitable secure app. The moving image can be displayed with surrounding text and images, to including branding and details of the voucher and promotion readable by the user.
The images and codes used here to describe the concept are barcodes, however, the same principles can be used to prevent other machine-readable images being duplicated, such as OR codes. The fake code images or pseudo-readable images used to obfuscate the genuine cade image should ideally be indistinguishable from genuine code images to the human eye, or at least to a significant proportion of the population, although the system would work with any images that don't comprise a coded image interspersed with a genuine code image, and the non-readable images could comprise shapes, photographs, black images, plain areas of colour, text, randomly generated images etc. Typically, the non-readable images will be pseudo-readable images as previously defined, having the same or similar appearance (e.g. the proportions and dimensions and general layout) as the genuine image.
For a typical barcode, the pseudo-readable image will comprise parallel lines of a contrasting colour or intensity to a background, the lines being of varying widths (for example, a random distribution of lines of two thicknesses), as for a genuine barcode. A fake OR code, for example, could include orientation markers, but they could be distributed in an unreadable configuration, or the area reserved for encoded data and error correction could be an unreadable black and white pattern.
It will be noted that the system described herein can be used in conjunction with a physical hard copy of a barcode. It also does not require any modification of existing scanning systems.
The setting for use of the moving image is described as being for vouchers for a store, for example this could be a supermarket. However, it will be realised that this system could be used for any situation where unauthorised duplication of barcodes or other machine-readable codes is a problem.
In this specification an apparatus/method/product "comprising" certain features is intended to be interpreted as meaning that it includes those features, but that it does not exclude the presence of other features.
Many variations are possible without departing from the scope of the present invention as defined in the appended claims.