We propose a privacy-preserving machine learning scheme with encryption-then-compression (EtC) images, where EtC images are images encrypted by using a block-based encryption method proposed for EtC systems with JPEG compression. In this paper, a novel property of EtC images is first discussed, although EtC ones was already shown to be compressible as a property. The novel property allows us to directly apply EtC images to machine learning algorithms non-specialized for computing encrypted data. In addition, the proposed scheme is demonstrated to provide no degradation in the performance of some typical machine learning algorithms including the support vector machine algorithm with kernel trick and random forests under the use of z-score normalization. A number of facial recognition experiments with are carried out to confirm the effectiveness of the proposed scheme.

We present an extended quarter-plane lattice model for generating two-dimensional (2-D) autoregressive fields. This work is a generalization of the extended lattice filter of diagonal form (ELDF) developed by Ertuzun et al. The proposed model represents a wider class of 2-D AR fields than conventional lattice models. Several examples are presented to demonstrate the applicability of the proposed model. Furthermore, the proposed structure is compared with other conventional lattice filters based on the computation of their entropy values.

To measure temperature dependent complex permittivity of dielectric materials, a rectangular cavity resonator with a heating system has been developed. In the experiment, microwave power with the frequency of 2.45 GHz is applied to heat the dielectric material. In order to reduce the error of the complex permittivity of dielectric material obtained from the perturbation method, an electromagnetic (EM) field simulator is applied which uses the Transmission Line Modeling (TLM) method. The uniformity of the temperature is also discussed by the use of heat transfer equation which applies the results of TLM simulation. It is found from the results that the accurate temperature dependence of complex permittivity of the material can be obtained by the method presented here.

In this paper, we propose an adaptive predictive coding method based on image segmentation for lossless compression. MAR (Multiplicative Autoregressive) predictive coding is an efficient lossless compression scheme. Predictors of the MAR model can be adapted to changes in the local image statistics due to its local image processing. However, the performance of the MAR method is reduced when applied to images whose local statistics change within the block-by-block subdivided image. Furthermore, side-information such as prediction coefficients must be transmitted to the decoder with each block. In order to enhance the compression performance, we improve the MAR coding method by using image segmentation. The proposed MAR predictor can be adapted to the local statistics of the image efficiently at each pixel. Furthermore, less side-information need be transmitted compared with the conventional MAR method.

Communications in locally structured wireless networks, such as ad hoc networks, will play an important role for network services and applications in the ubiquitous networking environments of the future. Most of the time, however, mobile networks are used in much the same way as fixed networks: most of the time, mobile terminals merely access information which is stored in the backbone network. Applications based on the local exchange of information gathered or generated by mobile terminals will open up many novel possibilities. A new online storage system named CAOSS facilitates such exchange and constitutes a building block for various new applications. CAOSS is a server-less system that provides high availability of data in mobile and wireless network environments. We describe CAOSS and its application in a mobile video-information-sharing system named GT. Thanks to the good performance of CAOSS, the GT system gives users a convenient and easy way to share and accumulate video data in a wireless environment. We confirm that CAOSS has strong enough performance to deal with large volumes of video data.

Multiple-input multiple-output (MIMO) OFDM technique is an attractive solution to increase the spectrum efficiency for mobile transmission applications. However, high spatial correlation makes signal detection difficult in real outdoor environments, and thus various methods have been developed to improve the detection performance. An iterative low-density parity-check (LDPC) coded multiple-input multiple-output (MIMO) system is a promising method for solving this problem, and its performance has been analyzed theoretically. This paper proposes an iterative LDPC minimum mean square error with soft interference cancellation (LDPC-MMSE-SIC) receiver with a time de-interleaver in front of the MMSE detector and evaluates its performance by computer simulation using channel state information (CSI) acquired in real outdoor measurements. We show that the iterative detection and decoding system with time interleaving, which is long enough to cover a fading cycle, achieves excellent error rate performance in mobile LOS environments and outperforms an LDPC maximum likelihood detection (LDPC-MLD) receiver with the same error correction and interleaving.

The method for estimating propagation loss that classifies receiving points into multiple groups by focusing on the number of reflections and diffractions, and applies a separate statistical model to each group was extended from only 2.4 GHz band to both 2.4 GHz and 5 GHz band. The extended statistical model was created from received power measurements. First, an appropriate grouping method was investigated based on the fitting error of statistical model. Non-line-of-sight (NLOS) receiving points were grouped in order of points that a wave reflected one time reaches, points that a wave reflected two times reaches, and points that a wave diffracted one time reaches. Next, the effectiveness of the proposed method was verified by comparison with conventional statistical models (one-slope, dual-slope, multi-wall, partitioned) on three office floors that differ from the environment used to create the statistical model. The average NLOS estimation error for the three evaluation environments was 4.9 dB, demonstrating that the proposed method has accuracy equal to or better than that of conventional methods.

For future adaptive networking services, it will be effective if autonomous agents utilize widely spread information that continuously changes. A communication network should provide a framework for agents to utilize information in a large-scale network and should positively support communications among agents. This paper proposes EcAgent (Environment for Communities of Agents). EcAgent provides agents with access to communities that are loose collections of agents. It gathers global information about communications among all agents and makes the agents share it in communities. It makes possible adaptive networking services that use various types of information. We implemented the EcAgent execution environment and some experimental applications to make an adaptive networking service. We also implemented a simulation system to evaluate how accurately the library recognized the embedded relationships as communities. The results of experiments show that the proposed methods for community recognition work efficiently.

High frame rate (HFR) video is attracting strong interest since it is considered as a next step toward providing Ultra-High Definition video service. For instance, the Association of Radio Industries and Businesses (ARIB) standard, the latest broadcasting standard in Japan, defines a 120 fps broadcasting format. The standard stipulates temporally scalable coding and hierarchical transmission by MPEG Media Transport (MMT), in which the base layer and the enhancement layer are transmitted over different paths for flexible distribution. We have developed the first ever MMT transmitter/receiver module for 4K/120fps temporally scalable video. The module is equipped with a newly proposed encapsulation method of temporally scalable bitstreams with correct boundaries. It is also designed to be tolerant to severe network constraints, including packet loss, arrival timing offset, and delay jitter. We conducted a hierarchical transmission experiment for 4K/120fps temporally scalable video. The experiment demonstrated that the MMT module was successfully fabricated and capable of dealing with severe network constraints. Consequently, the module has excellent potential as a means to support HFR video distribution in various network situations.

A rectangular waveguide compatible NRD guide E-plane bandpass filter is proposed for 55 GHz band OFDM applications. The NRD guide E-plane bandpass filter is constructed by inserting a metal foil array in the E-plane of NRD guide. Simulation, fabrication, and handling of the filter are not difficult because each resonator is constructed by a couple of metal foils of a simple shape. A Chebyshev response 5-pole bandpass filter with a very narrow bandwidth of 550 MHz is designed and fabricated at 55 GHz band. Simulated and measured filter performances agree well with the design specifications. Insertion loss of the fabricated filter is found to be around 2.0 dB. Although temperature stability of the fabricated filter are found to be within manageable level, the adoption of cyclo olefin polymer can be one of solution for the temperature stability improvement.

We have developed the world's largest 16.7-inch hybrid in-cell touch panel. To realize the large sized in-cell touch panel, we applied a vertical Vcom system and low resistance sensor, which are JDI's original technologies. For glove touch function, we applied mutual bundled driving, which increases the signal intensity higher. The panel also has a low surface reflection, curved-shaped, and non-rectangular characteristics, which are particular requirements in the automotive market. The over 15-inch hybrid in-cell touch panel adheres to automotive quality requirements. We have also developed a force touch panel, which is a new human machine interface (HMI) based on a hybrid in-cell touch panel in automotive display. This study reports on the effect of the improvements on the in-plane variation of force touch and the value change of the force signal under different environment conditions. We also a introduce force touch implemented prototype.

We have successfully applied Generalized Multiprotocol Label Switching (GMPLS) architecture to the Wavelength Assignment Photonic Switching System (WAPS) to create an internet access system that can provide, between terminals, not only conventional best-effort type of IP packet forwarding, but also high-speed and Quality of Service (QoS)-guaranteed IP forwarding. In this paper the system architecture, system specifications, and system hardware/software implementations are described.

The present paper examines a two-dimensional (2-D) joint-process lattice estimator and its implementation for image restoration. The gradient adaptive lattice (GAL) algorithm is used to update the filter coefficients. The proposed adaptive lattice estimator can represent a wider class of 2-D FIR systems than the conventional 2-D lattice models. Furthermore, its structure possesses orthogonality between the backward prediction errors. These results in superior convergence and tracking properties versus the transversal and other 2-D adaptive lattice estimators. The validity of the proposed model for image restoration is evaluated through computer simulations. In the examples, the implementation of the proposed lattice estimator as 2-D adaptive noise cancellator (ANC) and 2-D adaptive line enhancer (ALE) is considered.

In this paper, we propose a two-dimensional (2-D) least-squares lattice (LSL) algorithm for the general case of the autoregressive (AR) model with an asymmetric half-plane (AHP) coefficient support. The resulting LSL algorithm gives both order and space recursions for the 2-D deterministic normal equations. The size and shape of the coefficient support region of the proposed lattice filter can be chosen arbitrarily. Furthermore, the ordering of the support signal can be assigned arbitrarily. Finally, computer simulation for modeling a texture image is demonstrated to confirm the proposed model gives rapid convergence.

The Millimeter-wave Mobile Camera (MiMoCam) developed by NHK STRL uses millimeter-wave band (42 GHz/55 GHz) to transmit Hi-Vision TV picture with high quality and low latency. Multiple-input multiple-output (MIMO) technology which uses a number of antennas at both the transmitter and receiver can be adapted to use to transmit higher quality Hi-Vision TV picture. The camera was intended to be used in a studio environment where there is a high degree of multi-path, however there are also many requests for the MiMoCam to be used outdoor. This will present a different channel statistics where the camera will be operating in a near line-of-sight (LOS) environment without much reflected waves. We have conducted an outdoor transmission test and measured the outdoors transmission performance of the proposed MIMO system to clarify the possibility of using the MiMoCam in outdoor environment. This paper introduces the features of the MiMoCam system and the MIMO transmission technique used in the MiMoCam and presents the findings of this outdoor test. It was also confirmed that channel correlation of the MIMO propagation channels were suppressed by using orthogonally polarized waves and bit error rate (BER) characteristics with respect to the average receiving carrier-to-noise ratio (CNR) was improved. Finally, we could find the feasibility of the MiMoCam outdoor operation from these results.

Recently, flexible and lightweight optical devices are desired from the practical viewpoint. We demonstrated roll-to-roll type Anti Reflection (AR) film fabricated by layer-by-layer (LBL) adsorption process. When deposition time was 2.5 min and repeating cycle was 8 cycles, refractive index of LBL layer was 1.499 at 632 nm and thickness was 93.1 nm, which are almost the same as those of batch type LBL layer. The minimum reflectance was about 0.6% at 600 nm and transmittance was over 75% at visible region. However as compared with batch type, roll type AR film has lower reflectance and transmittance. This reason is that the flow of solution and rinse and quantity of rinse was smaller, a number of bathes of roll type was lower than that of batch type. Furthermore, comparing the deposition time and film speed, LBL layer was fabricated clearly long deposition time and slow film speed. The roll-to-roll film had a problem of peeling off during the deposition process. By increasing the contact area between film and guided roll, vertical pressure was decrease and friction force was decreased. Furthermore, as rotational speed of guided roll and film speed was decreased, LBL layer was not peeled by friction force between film and guided roll. Because rotational speed of guided rolls and films were almost same in the range of less than about 30 mm/min. There was the problem that polymer complexes were likely to appear on the substrate when the surface was dried during moving between solution and rinse bath. This phenomenon was observed during the roll-to-roll as well as batch process. The quality of roll-to-roll LBL process was depending on deposition time and film speed and drying at moving process critically compared with batch type. It is necessary to design the roll-to-roll machine with care: important points are deposition time and film speed, drying at moving process.

Self-paced e-learning provides much more freedom in time and locale than traditional education as well as diversity of learning contents and learning media and tools. However, its limitations must not be ignored. Lack of information on learners' states is a serious issue that can lead to severe problems, such as low learning efficiency, motivation loss, and even dropping out of e-learning. We have designed a novel e-learning support system that can visually observe learners' non-verbal behaviors and estimate their learning states and that can be easily integrated into practical e-learning environments. Three pairs of internal states closely related to learning performance, concentration-distraction, difficulty-ease, and interest-boredom, were selected as targets of recognition. In addition, we investigated the practical problem of estimating the learning states of a new learner whose characteristics are not known in advance. Experimental results show the potential of our system.

This paper proposes a unified coding algorithm for the lossless and near-lossless compression of still color images. The proposed unified color image coding scheme can control the Peak Signal-to-Noise Ratio (PSNR) of the reconstructed image while the level of distortion on the RGB plane is suppressed to within a preset magnitude. In order to control the PSNR, the distortion level is adaptively changed at each pixel. An adaptive quantizer to control the distortion is designed on the basis of psychovisual criteria. Finally, experiments on Super High Definition (SHD) images show the effectiveness of the proposed algorithm.

Directions and speeds of the motion of rain areas are estimated for each type of rain fronts, using time differences detected in the rain attenuation of the Ku-band satellite radio wave signals that have been measured at Osaka Electro-Communication University (OECU) in Neyagawa, Osaka, Research Institute of Sustainable Humanosphere (RISH) in Uji, Kyoto, and MU Observatory (MU) of Kyoto University in Shigaraki, Shiga, for the past five years since September 2002. These directions and speeds are shown to agree well with those directly obtained from the motion of rain fronts in the weather charts published by Japan Meteorological Agency. The rain area motion is found to have characteristic directions according to each rain type, such as cold and warm fronts or typhoon. A numerical estimate of the effects of site diversity techniques indicates that between two sites among the three locations (OECU, RISH, MU) separated by 20-50 km, the joint cumulative time percentages of rain attenuation become lower as the two sites are aligned along the directions of rain area motion. In such a case, compared with the ITU-R recommendations, the distance required between the two sites may be, on an average, reduced down to about 60-70% of the conventional predictions.