The nearest neighbor method is a simple and flexible scheme for the classification of data points in a vector space. It predicts a class label of an unseen data point using a majority rule for the labels of known data points inside a neighborhood of the unseen data point. Because it sometimes achieves good performance even for complicated problems, several derivatives of it have been studied. Among them, the discriminant adaptive nearest neighbor method is particularly worth revisiting to demonstrate its application. The main idea of this method is to adjust the neighbor metric of an unseen data point to the set of known data points before label prediction. It often improves the prediction, provided the neighbor metric is adjusted well. For statistical shape analysis, shape classification attracts attention because it is a vital topic in shape analysis. However, because a shape is generally expressed as a matrix, it is non-trivial to apply the discriminant adaptive nearest neighbor method to shape classification. Thus, in this study, we develop the discriminant adaptive nearest neighbor method to make it slightly more useful in shape classification. To achieve this development, a mixture model and optimization algorithm for shape clustering are incorporated into the method. Furthermore, we describe several helpful techniques for the initial guess of the model parameters in the optimization algorithm. Using several shape datasets, we demonstrated that our method is successful for shape classification.
Bungo TAGA Shiho MORIAI Kazumaro AOKI
In this paper, we present several cryptanalyses of Hierocrypt-L1 block cipher, which was selected as one of the CRYPTREC recommended ciphers in Japan in 2003. We present a differential attack and an impossible differential attack on 8 S-function layers in a related-key setting. We first show that there exist the key scheduling differential characteristics which always hold, then we search for differential paths for the data randomizing part with the minimum active S-boxes using the above key differentials. We also show that our impossible differential attack is a new type.
An on-channel repeater (OCR) performing simultaneous reception and transmission at the same frequency is beneficial to improve spectral efficiency and coverage. In an OCR, it is important to cancel the feedback interference caused by imperfect isolation between the transmit and receive antennas, and least mean square (LMS) based adaptive filters are commonly used for this purpose. In this paper, we analyze the performance of the LMS based adaptive feedback canceller in terms of its transient behavior and the steady-state mean square error (MSE). Through a theoretical analysis, we derive iterative equations to compute transient MSEs and provide a procedure to simply evaluate steady-state MSEs for the adaptive feedback canceller. Simulation results performed to verify the theoretical MSEs show good agreement between the proposed theoretical analysis and the empirical results.
Rong HUANG Palaiahnakote SHIVAKUMARA Yaokai FENG Seiichi UCHIDA
To handle the variety of scene characters, we propose a cooperative multiple-hypothesis framework which consists of an image operator set module, an Optical Character Recognition (OCR) module and an integration module. Multiple image operators activated by multiple parameters probe suspected character regions. The OCR module is then applied to each suspected region and returns multiple candidates with weight values for future integration. Without the aid of the heuristic rules which impose constraints on segmentation area, aspect ratio, color consistency, text line orientations, etc., the integration module automatically prunes the redundant detection/recognition and pads the missing detection/recognition. The proposed framework bridges the gap between scene character detection and recognition, in the sense that a practical OCR engine is effectively leveraged for result refinement. In addition, the proposed method achieves the detection and recognition at the character level, which enables dealing with special scenarios such as single character, text along arbitrary orientations or text along curves. We perform experiments on the benchmark ICDAR 2011 Robust Reading Competition dataset which includes a text localization task and a word recognition task. The quantitative results demonstrate that multiple hypotheses outperform a single hypothesis, and be comparable with state-of-the-art methods in terms of recall, precision, F-measure, character recognition rate, total edit distance and word recognition rate. Moreover, two additional experiments are conducted to confirm the simplicity of parameter setting in this proposal.
Yutaka KATSUYAMA Yoshinobu HOTTA Masako OMACHI Shinichiro OMACHI
Reducing the time complexity of character matching is critical to the development of efficient Japanese Optical Character Recognition (OCR) systems. To shorten the processing time, recognition is usually split into separate pre-classification and precise recognition stages. For high overall recognition performance, the pre-classification stage must both have very high classification accuracy and return only a small number of putative character categories for further processing. Furthermore, for any practical system, the speed of the pre-classification stage is also critical. The associative matching (AM) method has often been used for fast pre-classification because of its use of a hash table and reliance on just logical bit operations to select categories, both of which make it highly efficient. However, a certain level of redundancy exists in the hash table because it is constructed using only the minimum and maximum values of the data on each axis and therefore does not take account of the distribution of the data. We propose a novel method based on the AM method that satisfies the performance criteria described above but in a fraction of the time by modifying the hash table to reduce the range of each category of training characters. Furthermore, we show that our approach outperforms pre-classification by VQ clustering, ANN, LSH and AM in terms of classification accuracy, reducing the number of candidate categories and total processing time across an evaluation test set comprising 116,528 Japanese character images.
HyunMin SEUNG Jong-Dae LEE Chang-Hwan KIM Jea-Gun PARK
In summary, we successfully fabricated the nonvolatile hybrid polymer 4F2 memory-cell. It was based on bistable state, which was observed in PS layer that is containing a Ni nanocrystals capped with NiO tunneling barrier sandwiched by Al electrodes. The current conduction mechanism for polymer memory-cell was demonstrated by fitting the I-V curves. The electrons were charged and discharged on Ni nanocrystals by tunneling through the NiO tunneling barrier. In addition, the memory-cell showed a good and reproducible nonvolatile memory-cell characteristic. Its memory margin is about 1.410. The retention-time is more than 105 seconds and the endurance cycles of program-and-erase is more than 250 cycles. Furthermore, Thefore, polymer memory-cell would be good candidates for nonvolatile 4F2 cross-bar memory-cell.
Kiyoshi MORIMOTO Nobuyasu SUZUKI Kazuhiko YAMANAKA Masaaki YURI Janet MILLIEZ Xinbing LIU
This report describes a crystallization method we developed for amorphous (a)-Si film by using 405-nm laser diodes (LDs). The proposed method has been used to fabricate bottom gate (BG) microcrystalline (µc)-Si TFTs for the first time. A µc-Si film with high crystallinity was produced and high-performance BG µc-Si TFTs with a field effect mobility of 3.6 cm2/Vs and a current on/off ratio exceeding 108 were successfully demonstrated. To determine the advantages of a 405-nm wavelength, a heat flow simulation was performed with full consideration of light interference effects. Among commercially available solid-state lasers and LDs with wavelengths having relatively high optical absorption coefficients for a-Si, three (405, 445, and 532 nm) were used in the simulation for comparison. Results demonstrated that wavelength is a crucial factor for the uniformity, efficiency, and process margin in a-Si crystallization for BG µc-Si TFTs. The 405-nm wavelength had the best simulation results. In addition, the maximum temperature profile on the gate electrode through the simulation well explained the actual crystallinity distributions of the µc-Si films.
Jong-Dae LEE Hyun-Min SEUNG Kyoung-Cheol KWON Jea-Gun PARK
In summary, we successfully developed the polymer nonvolatile 4F2 memory-cell. It was based on nonvolatile memory characteristics such as memory margin and retention time, which was observed in memory-cell embedded with Ag nanocrystals in PVK layer. The nonvolatile memory characteristics depend on the shape, distribution and isolation of Ag nanocrystals. Accordingly, the thickness of Ag film has an important role in optimizing the Ag nanocrystals. Therefore, the polymer nonvolatile memory-cell is fabricated by appropriate thickness of film and need an improvement of interface between Ag nanocrystals and PVK for sufficient nonvolatile memory characteristics.
Akio FUJIYOSHI Masakazu SUZUKI
In this paper, we study the minimum spanning tree problem with label selection, that is, the problem of finding a minimum spanning tree of a vertex-labeled graph where the weight of each edge may vary depending on the selection of labels of vertices at both ends. The problem is especially important as the application to mathematical OCR. It is shown that the problem is NP-hard. However, for the application to mathematical OCR, it is sufficient to deal with only graphs with small tree-width. In this paper, a linear-time algorithm for series-parallel graphs is presented. Since the minimum spanning tree problem with label selection is closely related to the generalized minimum spanning tree problem, their relation is discussed.
Minoru MORI Minako SAWAKI Junji YAMATO
This paper describes an adaptive feature extraction method that exploits category-specific information to overcome both image degradation and deformation in character recognition. When recognizing multiple fonts, geometric features such as directional information of strokes are often used but they are weak against the deformation and degradation that appear in videos or natural scenes. To tackle these problems, the proposed method estimates the degree of deformation and degradation of an input pattern by comparing the input pattern and the template of each category as category-specific information. This estimation enables us to compensate the aspect ratio associated with shape and the degradation in feature values and so obtain higher recognition accuracy. Recognition experiments using characters extracted from videos show that the proposed method is superior to the conventional alternatives in resisting deformation and degradation.
Jinzhong YU Qiming WANG Buwen CHENG Saowu CHEN Yuhua ZUO
Si-based photonic materials and devices, including SiGe/Si quantum structures, SOI and InGaAs bonded on Si, PL of Si nanocrystals, SOI photonic crystal filter, Si based RCE (Resonant Cavity Enhanced) photodiodes, SOI TO (thermai-optical) switch matrix were investigated in Institute of Semiconductors, Chinese Academy of Sciences. The main results in recent years are presented in the paper. The mechanism of PL from Si NCs embedded in SiO2 matrix was studied, a greater contribution of the interface state recombination (PL peak in 850~900 nm) is associated with larger Si NCs and higher interface state density. Ge dots with density of order of 1011 cm-2 were obtained by UHV/CVD growth and 193 nm excimer laser annealing. SOI photonic crystal filter with resonant wavelength of 1598 nm and Q factor of 1140 was designed and made. Si based hybrid InGaAs RCE PD with η of 34.4% and FWHM of 27 nm were achieved by MOCVD growth and bonding technology between InGaAs epitaxial and Si wafers. A 1616 SOI optical switch matrix were designed and made. A new current driving circuit was used to improve the response speed of a 44 SOI rearrangeable nonblocking TO switch matrix, rising and falling time is 970 and 750 ns, respectively.
The historical review of Taiwan's researching activities on the features of PECVD grown SiOx are also included to realize the performance of Si nanocrystal based MOSLED made by such a Si-rich SiOx film with embedded Si nanocrystals on conventional Si substrate. A surface nano-roughened Si substrate with interfacial Si nano-pyramids at SiOx/Si interface are also reviewed, which provide the capabilities of enhancing the surface roughness induced total-internal-reflection relaxation and the Fowler-Nordheim tunneling based carrier injection. These structures enable the light emission and extraction from a metal-SiOx-Si MOSLED.
We investigate the enhancement of the optical nonlinearity and the limit of the improvement of the response speed in CdSxSe1-x microcrystallites by measuring the effective optical nonlinear cross section (σeff), the energy decay time (T1) and the dephasing time in two kinds of semiconductor microcrystallites of CdS0.12Se0.8 microcrystallites embedded in alkaline multi-component glasses (CdSSeMs) and CdSe microcrystallites embedded in SiO2 thin film (CdSeMs). As the average radius of CdSSeMs decreases from 10 to 1 nm, the values of σeff and T1 gradually change from 2.610-16 to 1.110-16 cm2 and from dozens picoseconds to 4 psec, respectively. The size dependence of CdSSEMs shows that the energy level structure in the microcrystallite with a radius of less than a few nanometers is a two-level system, in which σeff is proportional to T2. The carrier recombination time (τ) of CdSSeMs with the average radius of 1 nm is estimated to 2 psec. As the average radius of a CdS0.12Se0.8 microcrystallite decreases from 9 to 3 nm, the values of T2 gradually change from 640 to 230 fsec at 18 K, respectively. The size and temperature dependences of T2 for the CdSSeMs show that there is the discrepancy between the theory and the measured T2. The discrepancy showes the presence of the acoustic-phonon-assisted relaxation processes other than the pure-dephasing processes. It is indicated that T2 becomes long by reducing the excessive acoustic-phonon-assisted relaxation processes, and that the longer T2 might enhance σeff. We investigate the enhancement of σeff in CdSeMs by making T2 longer. The τ, σeff, and T2 of CdSeM an average radius of 3 nm are 40 psec, 4.510-15 cm2, and 150 fsec at room temperature. The σeff is ten times as large as that of CdSSeM sample at the same average radius and the enhancement of σeff can be considered to be caused by the longer T2.
Katsunori MAKIHARA Yoshihiro OKAMOTO Hideki MURAKAMI Seiichiro HIGASHI Seiichi MIYAZAKI
Hydrogenated germanium films were fabricated in the thickness range of 7-98 nm on SiO2 at 150 by an rf glow discharge decomposition of 0.25% GeH4 diluted with H2, and the nucleation and growth of Ge nanocrystallites were measured from topographic and current images simultaneously taken by a conductive AFM probe after Cr contact formation on films so prepared. We have demonstrated that current images show fine grains in comparison with topographic images and the lateral evolution of the Ge grains with progressive film growth. The contrast in current images can be interpreted in terms of the difference in electron concentration between nanocrystalline grains and their boundaries.
Several 2 nm seed layers were sputtered to increase coercivity (Hc) and anisotropy (Ku) of CoCrPt/Ti perpendicular recording media. Among them 2 nm Ag seed layer was very effective to increase Hc of (Co78Cr22)100-xPtx/Ti (x = 14, 20). However, the effect was more pronounced when (Co78Cr22)100-xPtx/Ti became thinner. In addition α[=4π(dM/dH)Hc] decreased when the Ag layer was used. The film thickness below which the seed Ag layer was effective was reduced with decreasing Pt content. However, the Ag seed layer did not promote (0002) texture of Ti and CoCrPt layers. Domain size was reduced when the Ag seed layer was used. The effects of Ag seed layer are thought to be due to change of exchange constant of the grains, for which the grain boundary plays an important role. Effects of film thickness and Pt content can also be explained successfully by the variation of exchange constant due to grain boundary. Some experimental evidence as well as crude mode for exchange constant variation are given.
Manabu OHTA Atsuhiro TAKASU Jun ADACHI
Optical Character Reader (OCR) incorrect recognition is a serious problem when searching for OCR-scanned documents in databases such as digital libraries. In order to reduce costs, this paper proposes fuzzy retrieval methods for English text containing errors in the recognized text without correcting the errors manually. The proposed methods generate multiple search terms for each input query term based on probabilistic automata which reflect both error-occurrence probabilities and character-connection probabilities. Experimental results of test-set retrieval indicate that one of the proposed methods improves the recall rate from 95.96% to 98.15% at the cost of a decrease in precision from 100.00% to 96.01% with 20 expanded search terms.
Access control involves a check to see if a user has an access right to a resource and then a decision is made as to whether his/her access to the resource is to be allowed or denied. Typical access control models are the Discretionary Access Control Model, the Mandatory Access Control Model, and the Role-Based Access Control Model. Today, the Role-Based Access Control Model has become popular and is recognized as an effective method. However, until now, the Role-Based Access Control Model was adequate only for bureaucracy organizations, in which some roles are standardized and organizational hierarchy is stable. Team-Based Access Control models that were designed for team-based organizations have been proposed, but they do not reflect some features of an adhocracy organization, which are organic, temporary, not standardized, changeable, and obscure in terms of hierarchical relationship, such as a Task Force Team in the company. This study shows the characteristics of an adhocracy organization that differ from the existing bureaucracy organization, and then shows why existing access control models have caused some problems. Finally, a revised Role-Based Access Control model is proposed to solve those problems and is analyzed according to main evaluation standards.
Yukio HONDA Yoshiyuki HIRAYAMA Atsushi KIKUKAWA Masaaki FUTAMOTO
The effects of an intermediate layer of carbon on the structural and magnetic properties of a CoCrPtTa recording layer were investigated in double-layered perpendicular magnetic recording media with either amorphous CoTaZr or crystalline FeAlSi as soft magnetic backlayers. Introducing a thin layer of carbon enhanced the perpendicular magnetic anisotropy with both soft magnetic backlayers. This result suggests that the introduction of a non-magnetic intermediate layer is useful in improving the basic magnetic properties of the CoCr-alloy recording layer even when an amorphous soft magnetic backlayers is used.
Masaaki FUTAMOTO Yoshiyuki HIRAYAMA Nobuyuki INABA Yukio HONDA Atsushi KIKUKAWA
Microstructures of CoCr-alloy thin film media were investigated by cross-sectional transmission electron microscopy focussing on the initial growth region of the magnetic layer grown on nonmagnetic underlayers. An introduction of nonmagnetic hcp-CoCrRu layer between an hcp-CoCrPt recording layer and an hcp- or a bcc-underlayer improved the crystallographic quality of the initial growth region. Sharp compositional distributions of alloying elements at the interfaces of a CoCrPt/CoCrRu/CrTi perpendicular medium and a CoCrPt/CoCrRu/CrTi longitudinal medium were respectively confirmed by electron energy loss spectroscopy employing a finely focussed electron beam. Coercivity and squareness of the thin film media increased by realizing good hetero-epitaxy between the nonmagnetic and the magnetic hcp-layers.
Large amounts of color-printed documents are published now everyday. Some OCR approaches of color-printed document images are provided, but they cannot normally work if the input images skew. In the past years, many algorithms are provided to detect the skew of monochrome document images but none of them process color-printed document images. All of these methods assume that text is printed in black on a white background and cannot be applied to detect skew in color-printed document images. In this paper, we propose an algorithm to detect the skew angle of a color-printed document image and reconstruct it. Our approach first determines variation of color-transition count at each angle (from -45 to +45) and the angle of maximal variation is regarded as the skew angle. Then, a scanning-line model reconstructs the image. We test 100 color-printed document images of various kinds and get good results (93 succeed and 7 fail). The average processing time of A4 size image is 2.76 seconds and the reconstruction time is 3.97 seconds on a Pentium III 733 PC.