In this paper, an adaptive block-based text line extraction algorithm is proposed. Three global and two local parameters are defined to adapt the method to various handwritings in different languages. A document image is segmented into several overlapping blocks. The skew of each block is estimated. Text block is de-skewed by using the estimated skew angle. Text regions are detected in the de-skewed text block. A number of data points are extracted from the detected text regions in each block. These data points are used to estimate the paths of text lines. By thinning the background of the image including text line paths, text line boundaries or separators are estimated. Furthermore, an algorithm is proposed to assign to the extracted text lines the connected components which have intersections with the estimated separators. Extensive experiments on different standard datasets in various languages demonstrate that the proposed algorithm outperforms previous methods.
Fatemeh ABRISHAMIAN Katsumi MORISHITA
The adjustable range on post-fabrication resonance wavelength trimming of long-period fiber gratings was broadened toward the blue side, and the mechanisms of the resonance wavelength shifts caused by heating were investigated. It can be concluded that the glass structure relaxes more slowly than the residual stress with decreasing heating temperature and the blue shift caused by the residual stress relaxation appears more strongly at the early stage of heating. The blue shift of 41 nm was obtained by heating a long-period grating at 600 for 3500 minutes. The changes of the index difference inducing the wavelength shifts of -41 nm and 35 nm were estimated at about -1.210-4 and +1.0 10-4 by numerical analysis, respectively.
Shoji KANEKO Masashi FUSHIKI Masayuki NAKANO Yoji KISHI
Multi-site MIMO (Multiple Input Multiple Output) is a key technology that will enable next generation cellular networks to achieve high throughput in cell edge areas. However, a multi-site single-user MIMO system is subject to performance degradation in terms of cell throughput due to the expense of additional assignments of radio resources to cell edge user equipment. This paper presents a BS-cooperation scheduling scheme for a multi-site single-user MIMO cellular system. The proposed BS-cooperation scheduling scheme aims to maintain cell throughput while improving cell edge user throughput. The proposed scheme employs two policies with respect to the assignment of radio resource to the user equipment with multi-site connection. One is to control the opportunities for radio resource assignment to user equipment with a multi-site connection to avoid the excessive assignment of radio resources and to maintain cell throughput. The other policy governs the decision as to whether the user equipment operates with a multi-site connection or not, making it possible for the multi-site connection to contribute to the improvement in user throughput in the cell edge areas. The simulation results show that the proposed scheme is effective from the perspective of both cell throughput and cell edge user throughput.
Tianruo ZHANG Chen LIU Minghui WANG Satoshi GOTO
This paper proposes a region-of-interest (ROI) based H.264 encoder and the VLSI architecture of the ROI detection algorithm. In ROI based video coding system, pre-processing unit to detect ROI should only introduce low computational complexity overhead due to the low power requirement. The Macroblocks (MBs) in ROIs are detected sequentially in the same order of H.264 encoding to satisfy the MB level pipelining of ROI detector and H.264 encoder. ROI detection is performed in a novel estimation-and-verification process with an ROI contour template. Proposed architecture can be configured to detect either single ROI or multiple ROIs in each frame and the throughput of single detection mode is 5.5 times of multiple detection mode. 98.01% and 97.89% of MBs in ROIs can be detected in single and multiple detection modes respectively. Hardware cost of proposed architecture is only 4.68 k gates. Detection speed is 753 fps for CIF format video at the operation frequency of 200 MHz in multiple detection mode with power consumption of 0.47 mW. Compared with previous fast ROI detection algorithms for video coding application, the proposed architecture obtains more accurate and smaller ROI. Therefore, more efficient ROI based computation complexity and compression efficiency optimization can be implemented in H.264 encoder.
Yiqing HUANG Xiaocong JIN Jin ZHOU Jia SU Takeshi IKENAGA
One high profile intra predictor generation engine is proposed in this paper. Firstly, hardware level algorithm optimization for intra 88 (I8MB) mode is introduced. The original candidate pixels for generating prediction samples of I8MB are replaced with boundary pixels of intra 44 (I4MB) blocks. Based on this adoption, full data reuse between predictors of I4MB and filtered samples of I8MB can be achieved with almost no quality loss. Secondly, one lossless two-44-block based parallel predictor generation flow is proposed. The original predictor generation flow is optimized from 16 stages to 10 stages for I4MB and Intra 1616 (I16MB), which saves 37.5% processing cycles. For I8MB, similar methodology with different processing order of 44 scaled blocks is introduced. Thirdly, fully utilized hardwired engines for I4MB, I16MB and I8MB are proposed in this paper. Except DC (direct current) and plane modes, full data reuse among all intra modes of high profile can be achieved. Fourthly, for DC mode, one combined predictor generation process is introduced and predictor generation of I16MB's DC mode is merged into the process of I4MB's DC mode. Moreover, by configuring proposed hardwired engines, predictor generation of I16MB's plane mode and chrominance plane mode can be accomplished with only 50% cycles of original design. Totally, when compared with original full-mode design and latest dynamic mode reused design, the proposed predictor generation engine can achieve 89.5% and 73.2% saving of processing cycles, respectively. Synthesized by TSMC 0.18 µm technology under worst work conditions (1.62 V, 125°C), with 380 MHz and 37.2 k gates, the proposed design can handle real-time high profile intra predictor generation of Super Hi-Vision 4 k4 k@60 fps. The maximum work frequency of our design under worst condition is 468 MHz.
Yoshimitsu TAKAMATSU Ryuichi FUJIMOTO Tsuyoshi SEKINE Takaya YASUDA Mitsumasa NAKAMURA Takuya HIRAKAWA Masato ISHII Motohiko HAYASHI Hiroya ITO Yoko WADA Teruo IMAYAMA Tatsuro OOMOTO Yosuke OGASAWARA Masaki NISHIKAWA Yoshihiro YOSHIDA Kenji YOSHIOKA Shigehito SAIGUSA Hiroshi YOSHIDA Nobuyuki ITOH
This paper presents a single-chip RF tuner/OFDM demodulator for a mobile digital TV application called “1-segment broadcasting.” To achieve required performances for the single-chip receiver, a tunable technique for a low-noise amplifier (LNA) and spurious suppression techniques are proposed in this paper. Firstly, to receive all channels from 470 MHz to 770 MHz and to relax distortion characteristics of following circuit blocks such as an RF variable-gain amplifier and a mixer, a tunable technique for the LNA is proposed. Then, to improve the sensitivity, spurious signal suppression techniques are also proposed. The single-chip receiver using the proposed techniques is fabricated in 90 nm CMOS technology and total die size is 3.26 mm 3.26 mm. Using the tunable LNA and suppressing undesired spurious signals, the sensitivities of less than -98.6 dBm are achieved for all the channels.
Jiangtao SUN Qing LIU Yong-Ju SUH Takayuki SHIBATA Toshihiko YOSHIMASU
A balanced push-push frequency doubler has been demonstrated in 0.25-µm SOI (Silicon on Insulator) SiGe BiCMOS technology operating from 22 GHz to 29 GHz with high fundamental frequency suppression and high conversion gain. A series LC resonator circuit is connected in parallel with the differential outputs of the doubler core circuit. The LC resonator is effective to improve the fundamental frequency suppression. In addition, the LC resonator works as a matching circuit between the output of the doubler core and the input of the output buffer amplifier, which increases the conversion gain of the whole circuit. A measured fundamental frequency suppression of greater than 46 dBc is achieved at an input power of -10 dBm in the output frequency band of 22-29 GHz. Moreover, maximum fundamental frequency suppression of 66 dBc is achieved at an input frequency of 13 GHz and an input power of -10 dBm. The frequency doubler works at a supply voltage of 3.3 V.
Ryuichi FUJIMOTO Kyoya TAKANO Mizuki MOTOYOSHI Uroschanit YODPRASIT Minoru FUJISHIMA
Device modeling techniques for high-frequency circuits operating at over 100 GHz are presented. We have proposed the bond-based design as an accurate high-frequency circuit design method. Because layout parasitic extractions (LPE) are not required in the bond-based design, it can be applied high-frequency circuit design at over 100 GHz. However, customized device models are indispensable for the bond-based design. In this paper, device modeling techniques for high-frequency circuit design using the bond-based design are proposed. The customized device model for MOSFETs, transmission lines and pads are introduced. By using customized device models, the difference between the simulated and measured gains of an amplifier is improved to less than 0.6 dB at 120 GHz.
In this paper, we present a fault analysis of the original NTRU public key cryptosystem. The fault model in which we analyze the cipher is the one in which the attacker is assumed to be able to fault a small number of coefficients of the polynomial input to (or output from) the second step of the decryption process but cannot control the exact location of injected faults. For this specific original instantiation of the NTRU encryption system with parameters (N,p,q), our attack succeeds with probability≈ and when the number of faulted coefficients is upper bounded by t, it requires O((pN)t) polynomial inversions in Z/p Z[x]/(xN-1).
Toru SHIMIZU Kazutami ARIMOTO Osamu NISHII Sugako OTANI Hiroyuki KONDO
Various low power technologies have been developed and applied to LSIs from the point of device and circuit design. A lot more CPU cores as well as function IPs are integrated on a single chip LSI today. Therefore, not only the device and circuit low power technologies, but software power control technologies are becoming more important to reduce active power of application systems. This paper overviews the low power technologies and defines power management platform as a combination of hardware functions and software programming interface. This paper discusses importance of the power management platform and direction of its development.
This paper presents an improved Gini-Index algorithm to correct feature-selection bias in text classification. Gini-Index has been used as a split measure for choosing the most appropriate splitting attribute in decision tree. Recently, an improved Gini-Index algorithm for feature selection, designed for text categorization and based on Gini-Index theory, was introduced, and it has proved to be better than the other methods. However, we found that the Gini-Index still shows a feature selection bias in text classification, specifically for unbalanced datasets having a huge number of features. The feature selection bias of the Gini-Index in feature selection is shown in three ways: 1) the Gini values of low-frequency features are low (on purity measure) overall, irrespective of the distribution of features among classes, 2) for high-frequency features, the Gini values are always relatively high and 3) for specific features belonging to large classes, the Gini values are relatively lower than those belonging to small classes. Therefore, to correct that bias and improve feature selection in text classification using Gini-Index, we propose an improved Gini-Index (I-GI) algorithm with three reformulated Gini-Index expressions. In the present study, we used global dimensionality reduction (DR) and local DR to measure the goodness of features in feature selections. In experimental results for the I-GI algorithm, we obtained unbiased feature values and eliminated many irrelevant general features while retaining many specific features. Furthermore, we could improve the overall classification performances when we used the local DR method. The total averages of the classification performance were increased by 19.4 %, 15.9 %, 3.3 %, 2.8 % and 2.9 % (kNN) in Micro-F1, 14 %, 9.8 %, 9.2 %, 3.5 % and 4.3 % (SVM) in Micro-F1, 20 %, 16.9 %, 2.8 %, 3.6 % and 3.1 % (kNN) in Macro-F1, 16.3 %, 14 %, 7.1 %, 4.4 %, 6.3 % (SVM) in Macro-F1, compared with tf*idf, χ2, Information Gain, Odds Ratio and the existing Gini-Index methods according to each classifier.
Jaesun KIM Younghoon KIM Hyuk-Jae LEE
The excessive memory access required to perform motion compensation when decoding compressed video is one of the main limitations in improving the performance of an H.264/AVC decoder. This paper proposes an H.264/AVC decoder that employs three techniques to reduce external memory access events: efficient distribution of reference frame data, on-chip cache memory, and frame memory recompression. The distribution of reference frame data is optimized to reduce the number of row activations during SDRAM access. The novel cache organization is proposed to simplify tag comparisons and ease the access to consecutive 4×4 blocks. A recompression algorithm is modified to improve compression efficiency by using unused storage space in neighboring blocks as well as the correlation with the neighboring pixels stored in the cache. Experimental results show that the three techniques together reduce external memory access time by an average of 90%, which is 16% better than the improvements achieved by previous work. Efficiency of the frame memory recompression algorithm is improved with a 32×32 cache, resulting in a PSNR improvement of 0.371 dB. The H.264/AVC decoder with the three techniques is fabricated and implemented as an ASIC using 0.18 µm technology.
Dongwan HONG Jeehee YOON Jongkeun LEE Sanghyun PARK Jongil KIM
By converting the expression values of each sample into the corresponding rank values, the rank-based approach enables the direct integration of multiple microarray data produced by different laboratories and/or different techniques. In this study, we verify through statistical and experimental methods that informative genes can be extracted from multiple microarray data integrated by the rank-based approach (briefly, integrated rank-based microarray data). First, after showing that a nonparametric technique can be used effectively as a scoring metric for rank-based microarray data, we prove that the scoring results from integrated rank-based microarray data are statistically significant. Next, through experimental comparisons, we show that the informative genes from integrated rank-based microarray data are statistically more significant than those of single-microarray data. In addition, by comparing the lists of informative genes extracted from experimental data, we show that the rank-based data integration method extracts more significant genes than the z-score-based normalization technique or the rank products technique. Public cancer microarray data were used for our experiments and the marker genes list from the CGAP database was used to compare the extracted genes. The GO database and the GSEA method were also used to analyze the functionalities of the extracted genes.
Internet group-based application layer services such as the overlay networks and P2P systems can benefit from end-to-end network status information. An efficient and accurate bandwidth measurement technique plays an important role in acquiring this information. We propose an end-to-end bottleneck link capacity measurement technique that utilizes path signatures combined with graphical analyses. This feature reduces the probe overhead and decreases the convergence time. We used ns-2 simulations and actual Internet measurements, which resulted in a high level of accuracy and a short probe time with low overhead.
Jung-Sun UM Sung-Hyun HWANG Chang-Joo KIM Byung Jang JEONG
Wireless regional area network (WRAN) is intended to offer the fixed wireless access services using cognitive radio technology in the TV white space. Therefore, WRAN shall minimize the transmission power so that harmful interference is not imposed on the licensed users operating in the TV bands. In this paper, we propose a processing block that offers improvements in the SNR and diversity gain using the block to algebraically process two constellation symbols. Thus, the transmission power can be reduced by an amount equal to the gains. The simulation result shows that the proposed scheme has a better bit error performance than the transmission scheme defined in the IEEE 802.22 draft standard.
Xi YANG Shengliang PENG Pengcheng ZHU Hongyang CHEN Xiuying CAO
The sensing scheme based on the generalized likelihood ratio test (GLRT) technique has attracted a lot of research interest in the field of cognitive radios (CR). Although its potential advantages in detecting correlated primary signal have been illustrated in prior work, no theoretical analysis of the positive effects of the correlation has appeared in the literature. In this letter, we derive the theoretical false-alarm and detection probabilities of GLRT detector. The theoretical analysis shows that, in the low signal-to-noise ratio (SNR) region, the detector's performance can be improved by exploiting the high correlations between the primary signal samples. The conclusions of the analysis are verified by numerical simulation results.
This paper shows a fast estimation method of very low error rate of low-density parity-check (LDPC) codes. No analytical tool is available to evaluate performance of LDPC codes, and the traditional Monte Carlo simulation methods can not estimate the low error rate of LDPC codes due to the limitation of time. To conquer this problem, we propose another simulation method which is based on the optimal simulation probability density function (PDF). The proposed simulation PDF can also avoid the dependency between the simulation time and the number of dominant trapping sets, which is the problem of some fast simulation methods based on the error event simulation method. Additionally, we show some numerical examples to demonstrate the effectiveness of the proposed method. The simulation time of the proposed method is reduced to almost less than 1/10 of that of Cole et al.'s method under the condition of the same accuracy of the estimator.
In this paper, we present a new frequency identification technique using the recent methodology of compressive sensing and discrete prolate spheroidal sequences with optimal energy concentration. Using the bandpass form of discrete prolate spheroidal sequences as basis matrix in compressive sensing, compressive frequency sensing algorithm is presented. Simulation results are given to present the effectiveness of the proposed technique for application to detection of carrier-frequency type signal and recognition of wideband signal in communication.
Juinn-Dar HUANG Chia-I CHEN Yen-Ting LIN Wan-Ling HSU
In deep-submicron era, wire delay is becoming a bottleneck while pursuing even higher system clock speed. Several distributed register (DR) architectures have been proposed to cope with this problem by keeping most wires local. In this article, we propose a new resource-constrained communication synthesis algorithm for optimizing both inter-island connections (IICs) and latency targeting on distributed register-file microarchitecture (DRFM). The experimental results show that up to 24.7% and 12.7% reduction on IIC and latency can be achieved respectively as compared to the previous work.
Keisuke INOUE Mineo KANEKO Tsuyoshi IWAGAKI
For recent and future nanometer-technology VLSIs, static and dynamic delay variations become a serious problem. In many cases, the hold timing constraint, as well as the setup timing constraint, becomes critical for latching a correct signal under delay variations. While the timing violation due to the fail of the setup timing constraint can be fixed by tuning a clock frequency or using a delayed latch, the timing violation due to the fail of the hold timing constraint cannot be fixed by those methods in general. Our approach to delay variations (in particular, the hold timing constraint) proposed in this paper is a novel register assignment strategy in high-level synthesis, which guarantees safe clocking by Backward-Data-Direction (BDD) clocking. One of the drawbacks of the proposed register assignment is the increase in the number of required registers. After the formulation of this new register minimization problem, we prove NP-hardness of the problem, and then derive an integer linear programming formulation for the problem. The proposed method receives a scheduled data flow graph, and generates a datapath having (1) robustness against delay variations, which is ensured by BDD-based register assignment, and (2) the minimum possible number of registers. Experimental results show the effectiveness of the proposed method for some benchmark circuits.