Artifact-metrics is an automated method of authenticating artifacts based on their measurable intrinsic characteristics
such as microscopic random-patterns and the like which are emerging in their manufacturing process. They are very
difficult to copy and control even for legal manufacturers. This paper describes the feasibility of an artifact-metric
system using laser speckle patterns of the surface of security documents. Speckle patterns of the surface differ from each
object depending on their microscopic geometry; they can be authenticated automatically by some matching algorithm
that compares the input with the previously enrolled data. Our system consists of a laser diode and a CMOS camera
which are attached firmly to a rigid metal frame and a matching algorithm by correlation coefficient method. To improve
matching accuracy we investigate the influence of width and incident angle of the laser beam and aperture size and
shooting angle for the camera on the matching accuracy through matching tests using 100 of ID cards. As a result, we
found that aperture size of the camera and width of the laser are part of important factors for matching accuracy. Also,
we achieved equal error rate of 0.1% using data size of 16 square pixels for speckle patterns.
Artifact-metrics is an automated method of authenticating artifacts based on a measurable intrinsic characteristic.
Intrinsic characters, such as microscopic random-patterns made during the manufacturing process, are very difficult to
copy. A transmitted light image of the distribution can be used for artifact-metrics, since the fiber distribution of paper is
random. Little is known about the individuality of the transmitted light image although it is an important requirement for
intrinsic characteristic artifact-metrics. Measuring individuality requires that the intrinsic characteristic of each artifact
significantly differs, so having sufficient individuality can make an artifact-metric system highly resistant to brute force
attack. Here we investigate the influence of paper category, matching size of sample, and image-resolution on the
individuality of a transmitted light image of paper through a matching test using those images. More concretely, we
evaluate FMR/FNMR curves by calculating similarity scores with matches using correlation coefficients between pairs
of scanner input images, and the individuality of paper by way of estimated EER with probabilistic measure through a
matching method based on line segments, which can localize the influence of rotation gaps of a sample in the case of
large matching size. As a result, we found that the transmitted light image of paper has a sufficient individuality.
Steganography aims to make communication invisible by hiding genuine information in innocent objects. We have proposed the SMF steganography that enables to hide information into Standard MIDI Files (SMF). The SMF is widely used as a standard storage format of the data of Musical Instrumental Digital Interface (MIDI). Most of digital musical instruments and personal computers equip the MIDI. Our hiding method utilizes a redundancy of the description of note events (note on/off) in SMF. Some note events, which are performed simultaneously, are allowed as correct contents of SMF even if they are described in any order. Therefore, we can permute the order of such note events according to the embedded data without changing their sounds. To clarify the potential of the SMF steganography, we calculate the embeddable data size about over three hundred SMFs that are opened to the public on the Internet. As a result, the embedding rate, which is the percentage of embeddable data size to cover SMF size, shows about 1% on average, and the best case reaches about 4%. We also clarify the influence of the Quantize function (which adjusts the timing of note events) on the embeddable data size.
Potential threats caused by something like real fingers, which are called fake or artificial fingers, should be crucial for authentication based on fingerprint systems. Security evaluation against attacks using such artificial fingers has been rarely disclosed. Only in patent literature, measures, such as live and well detection, against fake fingers have been proposed. However, the providers of fingerprint systems usually do not mention whether or not these measures are actually implemented in emerging fingerprint systems for PCs or smart cards or portable terminals, which are expected to enhance the grade of personal authentication necessary for digital transactions. As researchers who are pursuing secure systems, we would like to discuss attacks using artificial fingers and conduct experimental research to clarify the reality. This paper reports that gummy fingers, namely artificial fingers that are easily made of cheap and readily available gelatin, were accepted by extremely high rates by 11 particular fingerprint devices with optical or capacitive sensors. We have used the molds, which we made by pressing our live fingers against them or by processing fingerprint images from prints on glass surfaces, etc. We describe how to make the molds, and then show that the gummy fingers, which are made with these molds, can fool the fingerprint devices.
Clone resistance of most of all anti-counterfeiting techniques depends upon secret identifiers or difficulty of reproduction. However, these kinds of techniques are no use if a counterfeiter has enough financial a power to work the secret out or toe overcome the difficulty. We have focused on, among others, random patterns which can be inevitably difficult to reproduce. Consequently, we have developed a document protection system which utilizes random patterns of magnetic micro-fibers in substrates of documents. We have also applied cryptography to the system. The system verifies and authenticates a stochastic feature using pre-stored template data and a digital signature. The stochastic feature and the digital signature respectively prevent 'dead copying' and counterfeiting/alteration of the recorded data such as a value of the card. Therefore security of the system depends upon difficulty both in reproduction of the stochastic feature and in analysis of the digital signature. We have evalut3ed the security of the system for several kinds of criminal attacks. Some criminals may try to counterfeit a stochastic feature by modeling on a genuine document. In our paper we discuss security evaluation of our document protection system against such kinds of criminal methods and also discus a countermeasure and its difficulty.
Diffraction gratings and liquid crystals are use4d as security features for document protection. We have studied the combination of these two features and have developed an optically variable device which is called 'CPLgram'. CPLgram is made of thin film of high polymer cholesteric liquid crystals which exhibit diffraction of light. The combination of diffraction gratings and cholesteric liquid crystals provides light diffraction, wavelength selectivity, color- shifting of reflection and circular polarization selectivity, for the deice. These characteristics make it vary difficult to counterfeit the devices. In our understanding, it is a new approach to combine these security features. The most striking characteristic of the combination is significant aspect of the combination is that customers can choose a device verification method from three types of methods depending on security requirements and/or cost demands; to examine the device (a) with the naked eye, (b) with a simple inspection tool, and (c) by verification equipment.
We have used the term 'clone' to refer to those things which are produced by methods such as counterfeiting, alteration, duplication or simulation. To satisfy the requirements of secure and low-cost techniques for preventing card fraud, we have recently developed a clone preventive system called 'FibeCrypt (Fiber Cryptosystem)' which utilizes physical characteristics. Each card has a canonical domain (i.e. a distinctive part), similar to fingerprints as the biometric measurement, made up of magnetic micro-fibers scattered randomly inside. We have applied cryptosystems to the system. FibeCrypt examines and authenticates the unique pattern of the canonical domain using pre-stored reference data and a digital signature. In our paper, the schemes and the features of this system are described in detail. The results of our examinations show the accuracy of authentication of the system. We conclude that this authentication technique which utilizes physical characteristics can be very effective for clone prevention in various fields.