Yutaka FUKUCHI Kouji HIRATA Joji MAEDA
In all-optical switches using the cascade of second harmonic generation and difference frequency mixing in periodically poled lithium niobate (PPLN) waveguide devices, walk-off between the fundamental and second harmonic pulses causes crosstalk between neighboring symbols, and limits the switching performance. In this paper, we numerically study retiming characteristics of all-optical switches that employ the PPLN waveguide devices with consideration for the effects of the crosstalk and for the input timing of the data and clock pulses. We find that the time offset between the data and clock pulses can control the timing jitter of the switched output; an appropriate offset can reduce the jitter while improving the switching efficiency.
Guanwen ZHANG Jien KATO Yu WANG Kenji MASE
In this paper, we propose a patch-wise learning based approach to deal with the multiple-shot people re-identification task. In the proposed approach, re-identification is formulated as a patch-wise set-to-set matching problem, with each patch set being matched using a specifically learned Mahalanobis distance metric. The proposed approach has two advantages: (1) a patch-wise representation that moderates the ambiguousness of a non-rigid matching problem (of human body) to an approximate rigid one (of body parts); (2) a patch-wise learning algorithm that enables more constraints to be included in the learning process and results in distance metrics of high quality. We evaluate the proposed approach on popular benchmark datasets and confirm its competitive performance compared to the state-of-the-art methods.
Jungnam BAE Saichandrateja RADHAPURAM Ikkyun JO Takao KIHARA Toshimasa MATSUOKA
We present a low-voltage digitally-controlled oscillator (DCO) with the third-order ΔΣ modulator utilized in the medical implant communication service (MICS) frequency band. An optimized DCO core operating in the subthreshold region is designed, based on the gm/ID methodology. Thermometer coder with the dynamic element matching and ΔΣ modulator are implemented for the frequency tuning. High frequency resolution is achieved by using the ΔΣ modulator. The ΔΣ-modulator-based LC-DCO implemented in a 130-nm CMOS technology has achieved the phase noise of -115.3 dBc/Hz at 200 kHz offset frequency with the tuning range of 382 MHz to 412 MHz for the MICS band. It consumes 700 µW from a 0.7-V supply voltage and has a high frequency resolution of 18 kHz.
Yuan WANG Wei SU Guangliang GUO Xing ZHANG
A novel dynamic element matching (DEM) method, called binary-tree random DEM (BTR-DEM), is presented for a Nyquist-rate current-steering digital-to-analog converter (DAC). By increasing or decreasing the number of unit current sources randomly at the same time, the BTR-DEM encoding reduces switch transition glitches. A 5-bit current-steering DAC with the BTR-DEM technique is implemented in a 65-nm CMOS technology. The measured spurious free dynamic range (SFDR) attains 42 dB for a sample rate of 100 MHz and shows little dependence on signal frequency.
Kouhei KASAGI Naoto OSHIMA Safumi SUZUKI Masahiro ASADA
In this study, we propose and fabricate an oscillator array composed of three resonant-tunneling-diode terahertz oscillators integrated with slot-coupled patch antennas, and which does not require a Si lens. We measure the radiation pattern for single and arrayed oscillator, and calculate the output power using the integration of the pattern. The output power of a single oscillator was found to be ~15 µW. However, using an array configuration, almost combined output power of ~55 µW was obtained.
Yuan LIANG Koji IWANO Koichi SHINODA
Most error correction interfaces for speech recognition applications on smartphones require the user to first mark an error region and choose the correct word from a candidate list. We propose a simple multimodal interface to make the process more efficient. We develop Long Context Match (LCM) to get candidates that complement the conventional word confusion network (WCN). Assuming that not only the preceding words but also the succeeding words of the error region are validated by users, we use such contexts to search higher-order n-grams corpora for matching word sequences. For this purpose, we also utilize the Web text data. Furthermore, we propose a combination of LCM and WCN (“LCM + WCN”) to provide users with candidate lists that are more relevant than those yielded by WCN alone. We compare our interface with the WCN-based interface on the Corpus of Spontaneous Japanese (CSJ). Our proposed “LCM + WCN” method improved the 1-best accuracy by 23%, improved the Mean Reciprocal Rank (MRR) by 28%, and our interface reduced the user's load by 12%.
In recent years, many variants of key point based image descriptors have been designed for the image matching, and they have achieved remarkable performances. However, to some images, local features appear to be inapplicable. Since theses images usually have many local changes around key points compared with a normal image, we define this special image category as the image with local changes (IL). An IL pair (ILP) refers to an image pair which contains a normal image and its IL. ILP usually loses local visual similarities between two images while still holding global visual similarity. When an IL is given as a query image, the purpose of this work is to match the corresponding ILP in a large scale image set. As a solution, we use a compressed HOG feature descriptor to extract global visual similarity. For the nearest neighbor search problem, we propose random projection indexed KD-tree forests (rKDFs) to match ILP efficiently instead of exhaustive linear search. rKDFs is built with large scale low-dimensional KD-trees. Each KD-tree is built in a random projection indexed subspace and contributes to the final result equally through a voting mechanism. We evaluated our method by a benchmark which contains 35,000 candidate images and 5,000 query images. The results show that our method is efficient for solving local-changes invariant image matching problems.
Miquel ESPI Masakiyo FUJIMOTO Tomohiro NAKATANI
We present a method for recognition of acoustic events in conversation scenarios where speech usually overlaps with other acoustic events. While speech is usually considered the most informative acoustic event in a conversation scene, it does not always contain all the information. Non-speech events, such as a door knock, steps, or a keyboard typing can reveal aspects of the scene that speakers miss or avoid to mention. Moreover, being able to robustly detect these events could further support speech enhancement and recognition systems by providing useful information cues about the surrounding scenarios and noise. In acoustic event detection, state-of-the-art techniques are typically based on derived features (e.g. MFCC, or Mel-filter-banks) which have successfully parameterized the spectrogram of speech but reduce resolution and detail when we are targeting other kinds of events. In this paper, we propose a method that learns features in an unsupervised manner from high-resolution spectrogram patches (considering a patch as a certain number of consecutive frame features stacked together), and integrates within the deep neural network framework to detect and classify acoustic events. Superiority over both previous works in the field, and similar approaches based on derived features, has been assessed by statical measures and evaluation with CHIL2007 corpus, an annotated database of seminar recordings.
Jeong-Min CHOI Robin SHRESTHA Sungho JEON Jong-Soo SEO
In this paper, we study a distributed time-reversal space-time block coded single-carrier (D-TR-STBC-SC) system for amplify-and-forward (AF) half-duplex relaying in frequency-selective Rayleigh fading channels. Under the imperfect channel estimation condition, we analyze the mean-square-error (MSE) performance of the optimal and channel-mismatched frequency domain minimum MSE (FD-MMSE) and least square (LS) equalization. Our analysis results show that, unlike the point-to-point communications, the channel-mismatched FD-MMSE equalization of D-TR-STBC-SC relaying network leads to the ceiling effect that the MSE increases as the signal-to-noise ratio (SNR) of relay-to-destination link increases. Decomposing the MSE, it is found that the primary cause of the ceiling effect is the source-to-destination link in the first time-slot, which makes the covariance matrix of noise vector ill-conditioned. In order to resolve the channel-mismatching problems in the equalization process, we develop optimum relay power control strategies by considering practical channel estimations, i.e., training-based LS and linear minimum MSE (LMMSE) channel estimations. It is shown that the optimum power control resolves the trade-off between MSE performance and relay power consumption, and improves the robustness against the channel-mismatching. Finally, we introduce a performance evaluation to demonstrate the performance of channel equalization combined with the proposed power controls in D-TR-STBC-SC relaying network.
Shuji SAKAI Koichi ITO Takafumi AOKI Takafumi WATANABE Hiroki UNTEN
Methods of window matching to estimate 3D points are the most serious factors affecting the accuracy, robustness, and computational cost of Multi-View Stereo (MVS) algorithms. Most existing MVS algorithms employ window matching based on Normalized Cross-Correlation (NCC) to estimate the depth of a 3D point. NCC-based window matching estimates the displacement between matching windows with sub-pixel accuracy by linear/cubic interpolation, which does not represent accurate sub-pixel values of matching windows. This paper proposes a technique of window matching that is very accurate using Phase-Only Correlation (POC) with geometric correction for MVS. The accurate sub-pixel displacement between two matching windows can be estimated by fitting the analytical correlation peak model of the POC function. The proposed method also corrects the geometric transformations of matching windows by taking into consideration the 3D shape of a target object. The use of the proposed geometric correction approach makes it possible to achieve accurate 3D reconstruction from multi-view images even for images with large transformations. The proposed method demonstrates more accurate 3D reconstruction from multi-view images than the conventional methods in a set of experiments.
Yesheng GAO Hui SHENG Kaizhi WANG Xingzhao LIU
A signal-model-based SAR image formation algorithm is proposed in this paper. A model is used to describe the received signal, and each scatterer can be characterized by a set of its parameters. Two parameter estimation methods via atomic decomposition are presented: (1) applying 1-D matching pursuit to azimuthal projection data; (2) applying 2-D matching pursuit to raw data. The estimated parameters are mapped to form a SAR image, and the mapping procedure can be implemented under application guidelines. This algorithm requires no prior information about the relative motion between the platform and the target. The Cramer-Rao bounds of parameter estimation are derived, and the root mean square errors of the estimates are close to the bounds. Experimental results are given to validate the algorithm and indicate its potential applications.
Houari SABIRIN Hiroshi SANKOH Sei NAITO
This paper presents an automatic method to track soccer players in soccer video recorded from a single camera where the occurrence of pan-tilt-zoom can take place. The automatic object tracking is intended to support texture extraction in a free viewpoint video authoring application for soccer video. To ensure that the identity of the tracked object can be correctly obtained, background segmentation is performed and automatically removes commercial billboards whenever it overlaps with the soccer player. Next, object tracking is performed by an attribute matching algorithm for all objects in the temporal domain to find and maintain the correlation of the detected objects. The attribute matching process finds the best match between two objects in different frames according to their pre-determined attributes: position, size, dominant color and motion information. Utilizing these attributes, the experimental results show that the tracking process can handle occlusion problems such as occlusion involving more than three objects and occluded objects with similar color and moving direction, as well as correctly identify objects in the presence of camera movements.
Wenjun ZHAO Takao ONOYE Tian SONG
In this paper, a specified hardware architecture of the Fast Mode Decision (FMD) algorithms presented by our previous work is proposed. This architecture is designed as an embedded mode dispatch module. On the basis of this module, some unnecessary modes can be skipped or the mode decision process can be terminated in advanced. In order to maintain a higher compatibility, the FMD algorithms are unitedly designed as an unique module that can be easily embedded into a common video codec for H.265/HEVC. The input and output interfaces between the proposed module and other parts of the codec are designed based on simple but effective protocol. Hardware synthesis results on FPGA demonstrate that the proposed architecture achieves a maximum frequency of about 193 MHz with less than 1% of the total resources consumed. Moreover, the proposed module can improve the overall throughput.
Wence ZHANG Yan NI Hong REN Ming CHEN Jianxin DAI
This letter presents performance analysis in the high signal-to-noise ratio (SNR) region for matched filter (MF) precoding in single cell Massive MIMO systems. The outage probability function is derived in closed form, and the data rate of each user is also given. We have also presented asymptotic analysis in terms of data rate for MF when the number of users and the number of antennas grow without bounds. The expressions of these analytical results are rather simple and are thus convenient for overall performance evaluation. The simulation results show that the analysis are very accurate.
Yusuf Nur WIJAYANTO Atsushi KANNO Hiroshi MURATA Tetsuya KAWANISHI Yasuyuki OKAMURA
A millimeter-wave radar receiver using a z-cut LiNbO3 optical modulator with orthogonal-gap-embedded patch-antennas on a low-k dielectric material is proposed. A millimeter-wave from a reflected radar signal can be received by the patch-antennas and converted directly to a lightwave through electro-optic modulation. A low-k dielectric material is used as a substrate for improving antenna gain. Additionally, an interaction length between millimeter-wave and lightwave electric fields becomes long. As a result, large modulation efficiency can be obtained, which is proportional to sensitivity of the millimeter-wave radar receiver. Optical millimeter-wave radar beam-forming can be obtained using the proposed device with meandering-gaps for controlling interaction between millimeter-wave and lightwave electric fields in electro-optic modulation. Analysis and experimentally demonstration of the proposed device are discussed and reported for 40GHz millimeter-wave bands. Optical millimeter-wave radar beam-forming in 2-D is also discussed.
Xiao Lei YUAN Lu GAN Hong Shu LIAO
We address a robust algorithm for the interference-plus-noise covariance matrix reconstruction (RA-INCMR) against random arbitrary steering vector mismatches (RASVMs) of the interferences, which lead to substantial degradation of the original INCMR beamformer performance. Firstly, using the worst-case performance optimization (WCPO) criteria, we model these RASVMs as uncertainty sets and then propose the RA-INCMR to obtain the robust INCM (RINCM) based on the Robust Capon Beamforming (RCB) algorithm. Finally, we substitute the RINCM back into the original WCPO beamformer problem for the sample covariance matrix to formulate the new RA-INCM-WCPO beamformer problem. Simulation results demonstrate that the performance of the proposed beamformer is much better than the original INCMR beamformer when there exist RASVMs, especially at low signal-to-noise ratio (SNR).
Xianshi JING Sheng SUN Lei ZHU
A miniaturized patch hybrid coupler with arbitrary power ratio and impedance transformation is proposed and designed by loading a pair of asymmetric cross slots on a squared patch resonator. To obtain the arbitrary power ratio and impedance transformation, the rectangular size of stepped slot ends should be well designed to be asymmetry and thus to obtain the different inductive loadings along two current paths. Theoretically, the equivalent transmission line model is first developed to understand the physical relationship between the patch and traditional branch-line hybrids. The matching/isolation and power ratio conditions are then derived at center frequency. By following a detailed design guideline, a prototype of the hybrid with 1:2 power ratio and 1:1.3 impedance transformation is designed and fabricated at 4.2 GHz. The measured results show a good agreement with simulated results, where the measured -10 dB impedance bandwidth achieves 18% and the bandwidth of 90°±6° phase difference is about 35% in a frequency range from 3.5 GHz to 5 GHz.
We present a matching method for 3D CAD assembly models consisting of multiple components. Here we need to distinguish the layouts and the materials of the components in addition to their shapes. A set of the feature quantities of an assembly model is extracted using projections from various angles. We show the effectiveness of our method experimentally for 3D CAD assembly models.
Yusuke SUZUKI Takayoshi SHOUDAI Tomoyuki UCHIDA Tetsuhiro MIYAHARA
A term tree pattern is a rooted ordered tree pattern which consists of ordered tree structures with edge labels and structured variables with labels. All variables with the same label in a term tree pattern can be simultaneously replaced with ordered trees isomorphic to the same rooted ordered tree. Then, a term tree pattern is suitable for representing structural features common to tree structured data such as XML documents on the web, the secondary structures of RNA in biology and parse trees describing grammatical structures of natural languages. Let $ott$ be the set of all term tree patterns which have one or more variables with the same label. Let $lott$ be the set of all term tree patterns t such that all variables in t have distinct labels. We remark that $lottsubsetneq ott$ holds. In this paper, we consider a problem, called Matching problem for term tree patterns, of deciding whether or not a given rooted ordered tree T is obtained from a given term tree pattern t by replacing variables in t with rooted ordered trees. We show that Matching problem for term tree patterns in $ott$ is NP-complete, by giving a reduction from the string pattern matching problem, which is NP-complete. Next, by giving operations on an interval, which is a set containing all integers between two given integers representing vertex identifiers, we propose an efficient algorithm for solving Matching problem for term tree patterns in $lottsubsetneq ott$. Then, we show that, when an ordered tree having N vertices and a term tree pattern $t in lott$ having n vertices are given, the proposed matching algorithm correctly solves this problem in O(nN) time.
Yizhong LIU Tian SONG Yiqi ZHUANG Takashi SHIMAMOTO Xiang LI
This paper proposes a novel greedy algorithm, called Creditability-Estimation based Matching Pursuit (CEMP), for the compressed sensing signal recovery. As proved in the algorithm of Stagewise Orthogonal Matching Pursuit (StOMP), two Gaussian distributions are followed by the matched filter coefficients corresponding to and without corresponding to the actual support set of the original sparse signal, respectively. Therefore, the selection for each support point is interpreted as a process of hypothesis testing, and the preliminarily selected support set is supposed to consist of rejected atoms. A hard threshold, which is controlled by an input parameter, is used to implement the rejection. Because the Type I error may happen during the hypothesis testing, not all the rejected atoms are creditable to be the true support points. The creditability of each preliminarily selected support point is evaluated by a well-designed built-in mechanism, and the several most creditable ones are adaptively selected into the final support set without being controlled by any extra external parameters. Moreover, the proposed CEMP does not necessitate the sparsity level to be a priori control parameter in operation. In order to verify the performance of the proposed algorithm, Gaussian and Pulse Amplitude Modulation sparse signals are measured in the noiseless and noisy cases, and the experiments of the compressed sensing signal recoveries by several greedy algorithms including CEMP are implemented. The simulation results show the proposed CEMP can achieve the best performances of the recovery accuracy and robustness as a whole. Besides, the experiment of the compressed sensing image recovery shows that CEMP can recover the image with the highest Peak Signal to Noise Ratio (PSNR) and the best visual quality.