Wenjie XIE De XU Yingjun TANG Geng CUI
Previous works show that the probabilistic Latent Semantic Analysis (pLSA) model is one of the best generative models for scene categorization and can obtain an acceptable classification accuracy. However, this method uses a certain number of topics to construct the final image representation. In such a way, it restricts the image description to one level of visual detail and cannot generate a higher accuracy rate. In order to solve this problem, we propose a novel generative model, which is referred to as multi-scale multi-level probabilistic Latent Semantic Analysis model (msml-pLSA). This method consists of two parts: multi-scale part, which extracts visual details from the image of diverse resolutions, and multi-level part, which concentrates multiple levels of topic representation to model scene. The msml-pLSA model allows for the description of fine and coarse local image detail in one framework. The proposed method is evaluated on the well-known scene classification dataset with 15 scene categories, and experimental results show that the proposed msml-pLSA model can improve the classification accuracy compared with the typical classification methods.
Mohammadreza KESHAVARZI Abbas MOHAMMADI Abdolali ABDIPOUR Fadhel M. GHANNOUCHI
The effect of DC offset on multi-input multi-output (MIMO) direct transceivers with adaptive modulation (AM) is discussed in this paper. A variable-rate variable-power (VRVP) AM system with perfect channel state information (P-CSI) at both the transmitter and receiver in a MIMO scenario is considered. The DC offset is modeled as a zero mean complex Gaussian distributed random variable. By this modeling of the DC offset, the analytical expression for degraded bit error rate (BER) is derived. To derive this analytical expression, we establish a reasonable approximation. The good agreement between the analytical and simulation results shows that the approximation is valid and confirms the accuracy of the analytical expressions. Moreover, an approach to improve the degraded BER in these systems is introduced. For this purpose, we introduce a design for AM MIMO systems that takes account of DC offset and its effectiveness is confirmed. Throughput analysis for the AM MIMO system in the presence of DC offset is presented in this paper too. An analytical expression for throughput is derived and approximated to a simpler equation. At last, throughput results are compared to the simulation outcomes.
Seulki LEE Jerald YOO Hoi-Jun YOO
A Real-time Capacitor Compensation (RCC) scheme is proposed for low power and continuous communication in the wearable inductive coupling transceiver. Since inductance values of wearable inductor vary dynamically with deterioration of its communication characteristics, the inductance value is monitored and its resonance frequency is adjusted by additive parallel/serial capacitors in real time. RLC Bridge for detection of the inductance variations and the Dual-edge Sampling Comparator for recognition of the variance direction are proposed. It is implemented in a 0.18 µm CMOS technology, and it occupies a 12.7 mm2 chip area. The proposed transceiver consumes only 426.6 µW at 4 Mbps data rate. The compensation time takes 4.78 µs, including 3 µs of detection and 1.78 µs for compensation process in worst case.
Xiaohan LIU Hideo MAKINO Kenichi MASE
The need for efficient movement and precise location of robots in intelligent robot control systems within complex buildings is becoming increasingly important. This paper proposes an indoor positioning and communication platform using Fluorescent Light Communication (FLC) employing a newly developed nine-channel receiver, and discusses a new location estimation method using FLC, that involves a simulation model and coordinate calculation formulae. A series of experiments is performed. Distance errors of less than 25 cm are achieved. The enhanced FLC system yields benefits such as greater precision and ease of use.
Aram KAWEWONG Sirinart TANGRUAMSUB Osamu HASEGAWA
A novel Position-Invariant Robust Feature, designated as PIRF, is presented to address the problem of highly dynamic scene recognition. The PIRF is obtained by identifying existing local features (i.e. SIFT) that have a wide baseline visibility within a place (one place contains more than one sequential images). These wide-baseline visible features are then represented as a single PIRF, which is computed as an average of all descriptors associated with the PIRF. Particularly, PIRFs are robust against highly dynamical changes in scene: a single PIRF can be matched correctly against many features from many dynamical images. This paper also describes an approach to using these features for scene recognition. Recognition proceeds by matching an individual PIRF to a set of features from test images, with subsequent majority voting to identify a place with the highest matched PIRF. The PIRF system is trained and tested on 2000+ outdoor omnidirectional images and on COLD datasets. Despite its simplicity, PIRF offers a markedly better rate of recognition for dynamic outdoor scenes (ca. 90%) than the use of other features. Additionally, a robot navigation system based on PIRF (PIRF-Nav) can outperform other incremental topological mapping methods in terms of time (70% less) and memory. The number of PIRFs can be reduced further to reduce the time while retaining high accuracy, which makes it suitable for long-term recognition and localization.
Dan-ni AI Xian-hua HAN Xiang RUAN Yen-wei CHEN
In this paper, we present a novel color independent components based SIFT descriptor (termed CIC-SIFT) for object/scene classification. We first learn an efficient color transformation matrix based on independent component analysis (ICA), which is adaptive to each category in a database. The ICA-based color transformation can enhance contrast between the objects and the background in an image. Then we compute CIC-SIFT descriptors over all three transformed color independent components. Since the ICA-based color transformation can boost the objects and suppress the background, the proposed CIC-SIFT can extract more effective and discriminative local features for object/scene classification. The comparison is performed among seven SIFT descriptors, and the experimental classification results show that our proposed CIC-SIFT is superior to other conventional SIFT descriptors.
Zhenpeng BIAN Ruohe YAO Fei LUO
A low-voltage class-AB CMOS output stage with a tunable quiescent current control circuit is presented. It is based on a complementary common source. The quiescent current is detected by a compact circuit and can be adjusted by means of a control current without need to modify the transistor dimensions. The minimum supply voltage can be down to one threshold voltage plus two saturation voltages. It is suitable to drive low resistive loads. Simulation results are provided that are in agreement with expected characteristics.
Junichi NAKAGAWA Masamichi NOGAMI Masaki NODA Naoki SUZUKI Satoshi YOSHIMA Hitoyuki TAGAMI
10G-EPON systems have attracted a great deal of attention as a way of exceeding to realize over 10 Gb/s for optical subscriber networking. Rapid burst-mode transmitting/receiving techniques are the key technologies enabling the burst-mode upstream transmission of 10G-EPON systems. In this paper, we have developed a OLT burst-mode 3R receiver incorporating a burst-mode AGC optical receiver and an 82.5 GS/s over-sampling burst-mode CDR and a ONU burst-mode transmitter with high launch power DFB-LD of 1.27 µm wavelength to fully compliant with IEEE802.3av 10G-EPON PR30 standards. The transmitting characteristics of a fast LD turn-on/off time of less than 6ns and a high launch power of more than +8.0 dBm, and the receiving characteristics of receiver sensitivity of -30.1 dBm and the upstream power budget of 38.1 dB are successfully achieved.
Giordano SPADACINI Sergio A. PIGNARI
This work presents a statistical model for the radiated susceptibility (RS) of an unshielded twisted-wire pair (TWP) running above ground, illuminated by a random electromagnetic field. The incident field is modeled as a superposition of elemental plane waves with random angular density, phase, and polarization. The statistical properties of both the differential-mode (DM) and the common-mode (CM) noise voltages induced across the terminal loads are derived and discussed.
Po-Hung CHEN Min-Chiao CHEN Chun-Lin KO Chung-Yu WU
A direct-conversion receiver integrated with the CMOS subharmonic frequency tripler (SFT) for V-band applications is designed, fabricated and measured using 0.13-µm CMOS technology. The receiver consists of a low-noise amplifier, a down-conversion mixer, an output buffer, and an SFT. A fully differential SFT is introduced to relax the requirements on the design of the frequency synthesizer. Thus, the operational frequency of the frequency synthesizer in the proposed receiver is only 20 GHz. The fabricated receiver has a maximum conversion gain of 19.4 dB, a minimum single-side band noise figure of 10.2 dB, the input-referred 1-dB compression point of -20 dBm and the input third order inter-modulation intercept point of -8.3 dB. It draws only 15.8 mA from a 1.2-V power supply with a total chip area of 0.794 mm0.794 mm. As a result, it is feasible to apply the proposed receiver in low-power wireless transceiver in the V-band applications.
Bag-of-Visual-Words representation has recently become popular for scene classification. However, learning the visual words in an unsupervised manner suffers from the problem when faced these patches with similar appearances corresponding to distinct semantic concepts. This paper proposes a novel supervised learning framework, which aims at taking full advantage of label information to address the problem. Specifically, the Gaussian Mixture Modeling (GMM) is firstly applied to obtain "semantic interpretation" of patches using scene labels. Each scene induces a probability density on the low-level visual features space, and patches are represented as vectors of posterior scene semantic concepts probabilities. And then the Information Bottleneck (IB) algorithm is introduce to cluster the patches into "visual words" via a supervised manner, from the perspective of semantic interpretations. Such operation can maximize the semantic information of the visual words. Once obtained the visual words, the appearing frequency of the corresponding visual words in a given image forms a histogram, which can be subsequently used in the scene categorization task via the Support Vector Machine (SVM) classifier. Experiments on a challenging dataset show that the proposed visual words better perform scene classification task than most existing methods.
Kenichi AGAWA Shinichiro ISHIZUKA Hideaki MAJIMA Hiroyuki KOBAYASHI Masayuki KOIZUMI Takeshi NAGANO Makoto ARAI Yutaka SHIMIZU Asuka MAKI Go URAKAWA Tadashi TERADA Nobuyuki ITOH Mototsugu HAMADA Fumie FUJII Tadamasa KATO Sadayuki YOSHITOMI Nobuaki OTSUKA
A 2.4 GHz 0.13 µm CMOS transceiver LSI, supporting Bluetooth V2.1+enhanced data rate (EDR) standard, has achieved a high reception sensitivity and high-quality transmission signals between -40 and +90. A low-IF receiver and direct-conversion transmitter architecture are employed. A temperature compensated receiver chain including a low-noise amplifier accomplishes a sensitivity of -90 dBm at frequency shift keying modulation even in the worst environmental condition. Design optimization of phase noise in a local oscillator and linearity of a power amplifier improves transmission signals and enables them to meet Bluetooth radio specifications. Fabrication in scaled 0.13 µm CMOS and operation at a low supply voltage of 1.5 V result in small area and low power consumption.
This paper proposes a method to generate alternative scenarios from a normal scenario written with a scenario language. This method includes (1) generation of alternative plans and (2) generation of alternative scenario by a user's selection of these plans. The proposed method enables users to decrease the omission of the possible alternative scenarios in the early stages of development. The method will be illustrated with some examples.
Masashi SUGIYAMA Ichiro TAKEUCHI Taiji SUZUKI Takafumi KANAMORI Hirotaka HACHIYA Daisuke OKANOHARA
Estimating the conditional mean of an input-output relation is the goal of regression. However, regression analysis is not sufficiently informative if the conditional distribution has multi-modality, is highly asymmetric, or contains heteroscedastic noise. In such scenarios, estimating the conditional distribution itself would be more useful. In this paper, we propose a novel method of conditional density estimation that is suitable for multi-dimensional continuous variables. The basic idea of the proposed method is to express the conditional density in terms of the density ratio and the ratio is directly estimated without going through density estimation. Experiments using benchmark and robot transition datasets illustrate the usefulness of the proposed approach.
Vishal V. KULKARNI Hiroki ISHIKURO Tadahiro KURODA
A CMOS wireless transceiver operating in the 14-18 GHz range is proposed. The receiver uses direct conversion architecture for demodulation with a fast carrier and symbol timing recovery scheme. The transmitter uses a PLL and an up-conversion mixer to generate BPSK modulated signal. A ring oscillator is used in the PLL to make faster switching for burst transmission obtaining high speed low power operation. The transceiver operation has been verified by system simulation while the transmitter test-chip was fabricated in 65 nm CMOS technology and verified with measured results. The transmitter generates a bi-phase modulated signal with a center frequency of 16 GHz at a maximum data rate of 4 Gb/s and consumes 61 mW of power. To the best knowledge of authors, this is lowest power consumption among the reported transmitters that operate over 1 Gb/s range. The transceiver is proposed for a target communication distance of 10 cm.
Takehiko UNO Satoru NOGE Kei KASAHARA
We report Ultraviolet (UV)-induced visible light luminescence in artificial-lattice thin films of ion-doped silica glass (silica superstructure thin films). The film was composed of periodic nanometer layers of germanium-doped silica (Ge:SiO2), titanium-doped silica (Ti:SiO2), and tin-doped silica (Sn:SiO2). The thickness of each layer was between 10 and 30 nm. Despite the small thickness of the film (few microns), a relatively bright luminescence of white light was observed, along with cathode-ray luminescence in the superstructure film. In addition, irradiation of the superstructure film with UV light led to light amplification by stimulated emission at 405 nm. The experimental results suggest the potential application of silica superstructure thin films as optical amplifiers.
Toshinari OGIWARA Jun-ichi TAKAHASHI Hitoshi KUMA Yuichiro KAWAMURA Toshihiro IWAKUMA Chishio HOSOKAWA
We carried out degradation analysis of a blue phosphorescent organic light emitting diode by both impedance spectroscopy and transient electroluminescence (EL) spectroscopy. The number of semicircles observed in the Cole-Cole plot of the modulus became three to two after the device was operated for 567 hours. Considering the effective layer thickness of the initial and degraded devices did not change by degradation and combining the analysis of the Bode-plot of the imaginary part of the modulus, the relaxation times of emission layer and hole-blocking with electron transport layers changed to nearly the same value by the increase of the resistance of emission layer. Decay time of transient EL of the initial device was coincident with that of the degraded one. These phenomena suggest that no phosphorescence quenching sites are generated in the degraded device, but the number of the emission sites decrease by degradation.
Xiaorong JING Tianqi ZHANG Zhengzhong ZHOU
Combining the sphere decoding (SD) algorithm and the sequential detection method, we propose an adaptive group detection (AGD) scheme based on the sort-descending QRD (S-D-QRD) for V-BLAST architectures over an i.i.d. Rayleigh flat fading channel. Simulation results show that the proposed scheme, which encompasses the SD algorithm and the sequential detection method as two extreme cases in a probability sense, can achieve a very flexible tradeoff between the detection performance and computational complexity by adjusting the group parameter.
Tang YINGJUN Xu DE Yang XU Liu QIFANG
We present a novel model named Integrated Latent Topic Model (ILTM), to learn and recognize natural scene category. Unlike previous work, which considered the discrepancy and common property separately among all categories, Our approach combines universal topics from all categories with specific topics from each category. As a result, the model is implemented to produce a few but specific topics and more generic topics among categories, and each category is represented in a different topics simplex, which correlates well with human scene understanding. We investigate the classification performance with variable scene category tasks. The experiments have shown our model outperforms latent-space methods with less training data.
A method for accurate scene segmentation using two kinds of directed graph obtained by object matching and audio features is proposed. Generally, in audiovisual materials, such as broadcast programs and movies, there are repeated appearances of similar shots that include frames of the same background, object or place, and such shots are included in a single scene. Many scene segmentation methods based on this idea have been proposed; however, since they use color information as visual features, they cannot provide accurate scene segmentation results if the color features change in different shots for which frames include the same object due to camera operations such as zooming and panning. In order to solve this problem, scene segmentation by the proposed method is realized by using two novel approaches. In the first approach, object matching is performed between two frames that are each included in different shots. By using these matching results, repeated appearances of shots for which frames include the same object can be successfully found and represented as a directed graph. The proposed method also generates another directed graph that represents the repeated appearances of shots with similar audio features in the second approach. By combined use of these two directed graphs, degradation of scene segmentation accuracy, which results from using only one kind of graph, can be avoided in the proposed method and thereby accurate scene segmentation can be realized. Experimental results performed by applying the proposed method to actual broadcast programs are shown to verify the effectiveness of the proposed method.