Multi-level divide-and-conquer (MDC) is a generalized divide-and-conquer technique, which consists of more than one division step organized hierarchically. In this paper, we investigate the paradigm of the MDC and show that it is an efficient technique for designing parallel algorithms. The following parallel algorithms are used for studying the MDC: finding the convex hull of discs, finding the upper envelope of line segments, finding the farthest neighbors of a convex polygon and finding all the row maxima of a totally monotone matrix. The third and the fourth algorithms are newly presented. Our discussion is based on the EREW PRAM, but the methods discussed here can be applied to any parallel computation models.
He LI Yutaro IWAMOTO Xianhua HAN Lanfen LIN Akira FURUKAWA Shuzo KANASAKI Yen-Wei CHEN
Convolutional neural networks (CNNs) have become popular in medical image segmentation. The widely used deep CNNs are customized to extract multiple representative features for two-dimensional (2D) data, generally called 2D networks. However, 2D networks are inefficient in extracting three-dimensional (3D) spatial features from volumetric images. Although most 2D segmentation networks can be extended to 3D networks, the naively extended 3D methods are resource-intensive. In this paper, we propose an efficient and accurate network for fully automatic 3D segmentation. Specifically, we designed a 3D multiple-contextual extractor to capture rich global contextual dependencies from different feature levels. Then we leveraged an ROI-estimation strategy to crop the ROI bounding box. Meanwhile, we used a 3D ROI-attention module to improve the accuracy of in-region segmentation in the decoder path. Moreover, we used a hybrid Dice loss function to address the issues of class imbalance and blurry contour in medical images. By incorporating the above strategies, we realized a practical end-to-end 3D medical image segmentation with high efficiency and accuracy. To validate the 3D segmentation performance of our proposed method, we conducted extensive experiments on two datasets and demonstrated favorable results over the state-of-the-art methods.
Shiyu TENG Jiaqing LIU Yue HUANG Shurong CHAI Tomoko TATEYAMA Xinyin HUANG Lanfen LIN Yen-Wei CHEN
Depression is a prevalent mental disorder affecting a significant portion of the global population, leading to considerable disability and contributing to the overall burden of disease. Consequently, designing efficient and robust automated methods for depression detection has become imperative. Recently, deep learning methods, especially multimodal fusion methods, have been increasingly used in computer-aided depression detection. Importantly, individuals with depression and those without respond differently to various emotional stimuli, providing valuable information for detecting depression. Building on these observations, we propose an intra- and inter-emotional stimulus transformer-based fusion model to effectively extract depression-related features. The intra-emotional stimulus fusion framework aims to prioritize different modalities, capitalizing on their diversity and complementarity for depression detection. The inter-emotional stimulus model maps each emotional stimulus onto both invariant and specific subspaces using individual invariant and specific encoders. The emotional stimulus-invariant subspace facilitates efficient information sharing and integration across different emotional stimulus categories, while the emotional stimulus specific subspace seeks to enhance diversity and capture the distinct characteristics of individual emotional stimulus categories. Our proposed intra- and inter-emotional stimulus fusion model effectively integrates multimodal data under various emotional stimulus categories, providing a comprehensive representation that allows accurate task predictions in the context of depression detection. We evaluate the proposed model on the Chinese Soochow University students dataset, and the results outperform state-of-the-art models in terms of concordance correlation coefficient (CCC), root mean squared error (RMSE) and accuracy.
Xian-Hua HAN Yen-Wei CHEN Zensho NAKAO
We propose a robust edge detection method based on independent component analysis (ICA). It is known that most of the basis functions extracted from natural images by ICA are sparse and similar to localized and oriented receptive fields, and in the proposed edge detection method, a target image is first transformed by ICA basis functions and then the edges are detected or reconstructed with sparse components only. Furthermore, by applying a shrinkage algorithm to filter out the components of noise in the ICA domain, we can readily obtain the sparse components of the original image, resulting in a kind of robust edge detection even for a noisy image with a very low SN ratio. The efficiency of the proposed method is demonstrated by experiments with some natural images.
Guowei CHEN Kenichi ITOH Takuro SATO
Routing in Ad-hoc networks is unreliable due to the mobility of the nodes. Location-based routing protocols, unlike other protocols which rely on flooding, excel in network scalability. Furthermore, new location-based routing protocols, like, e.g. BLR [1] , IGF [2], & CBF [3] have been proposed, with the feature of not requiring beacons in MAC-layer, which improve more in terms of scalability. Such beaconless routing protocols can work efficiently in dense network areas. However, these protocols' algorithms have no ability to avoid from routing into sparse areas. In this article, historical signal strength has been added as a factor into the BLR algorithm, which avoids routing into sparse area, and consequently improves the global routing efficiency.
Shiann-Tsong SHEU Yen-Chieh CHENG Ping-Jung HSIEH Jung-Shyr WU Luwei CHEN
Wireless access in the vehicular environment (WAVE) architecture of intelligent transportation system (ITS) has been standardized in the IEEE 802.11p specification and it is going to be widely deployed in many roadway environments in order to provide prompt emergency information and internet services. A typical WAVE network consists of a number of WAVE devices, in which one is the road-side-unit (RSU) and the others are on-board-units (OBUs), and supports one control channel (CCH) and one or more service channels (SCH) for OBU access. The CCH is used to transport the emergency messages and service information of SCHs and the SCHs could be used to carry internet traffic and non-critical safety traffic of OBUs. However, the IEEE 802.11p contention-based medium access control protocol would suffer degraded transmission efficiency if the number of OBUs contending on an SCH is large. Moreover, synchronizing all WAVE devices to periodically and equally access the CCH and an SCH will waste as much as 50% of the channel resources of the SCH [1]. As a solution, we propose an efficiency-improvement scheme, namely the agent-based coordination (ABC) scheme, which improves the SCH throughput by means of electing one OBU to be the agent to schedule the other OBUs so that they obtain the access opportunities on one SCH and access the other SCH served by RSU in a contention-free manner. Based on the ABC scheme, three different scheduling and/or relaying strategies are further proposed and compared. Numerical results and simulation results confirm that the proposed ABC scheme significantly promotes the standard transmission efficiency.
Mike Shuo-Wei CHEN Robert W. BRODERSEN
This paper describes a system architecture along with signal processing technique which allows a reduction in the complexity of a 3.1-10.6 GHz Ultra-Wideband radio. The proposed system transmits passband pulses using a pulser and antenna, and the receiver front-end down-converts the signal frequency by subsampling, thus, requiring substantially less hardware than a traditional narrowband approach. However, the simplified receiver front end shows a high sensitivity to timing offset. By proposing an analytic signal processing technique, the vulnerability of timing offset is mitigated; furthermore, a time resolution finer than the sampling period is achieved, which is useful for locationing or ranging applications. Analysis and simulations of system specifications are also provided in this paper.
Yong HE Ji LI Xuanhong ZHOU Zewei CHEN Xin LIU
6DoF pose estimation from a monocular RGB image is a challenging but fundamental task. The methods based on unit direction vector-field representation and Hough voting strategy achieved state-of-the-art performance. Nevertheless, they apply the smooth l1 loss to learn the two elements of the unit vector separately, resulting in which is not taken into account that the prior distance between the pixel and the keypoint. While the positioning error is significantly affected by the prior distance. In this work, we propose a Prior Distance Augmented Loss (PDAL) to exploit the prior distance for more accurate vector-field representation. Furthermore, we propose a lightweight channel-level attention module for adaptive feature fusion. Embedding this Adaptive Fusion Attention Module (AFAM) into the U-Net, we build an Attention Voting Network to further improve the performance of our method. We conduct extensive experiments to demonstrate the effectiveness and performance improvement of our methods on the LINEMOD, OCCLUSION and YCB-Video datasets. Our experiments show that the proposed methods bring significant performance gains and outperform state-of-the-art RGB-based methods without any post-refinement.
Yen-Wei CHEN Zensho NAKAO Kouichi ARAKAKI Shinichi TAMURA
A genetic algorithm is presented for the blind-deconvolution problem of image restoration without any a priori information about object image or blurring function. The restoration problem is modeled as an optimization problem, whose cost function is to be minimized based on mechanics of natural selection and natural genetics. The applicability of GA for blind-deconvolution problem was demonstrated.
Xiang-Yan ZENG Yen-Wei CHEN Zensho NAKAO Hanqing LU
We propose a novel pixel pattern-based approach for texture classification, which is independent of the variance of illumination. Gray scale images are first transformed into pattern maps in which edges and lines, used for characterizing texture information, are classified by pattern matching. We employ principal component analysis (PCA) which is widely applied to feature extraction. We use the basis functions learned through PCA as templates for pattern matching. Using PCA pattern maps, the feature vector is comprised of the numbers of the pixels belonging to a specific pattern. The effectiveness of the new feature is demonstrated by applications to the image retrievals of the Brodatz texture database. Comparisons with multichannel and multiresolution features indicate that the new feature is quite time saving, free of the influence of illumination, and has comparable accuracy.
Yen-Wei CHEN Noriaki MIYANAGA Minoru UNEMOTO Masanobu YAMANAKA Tatsuhiko YAMANAKA Sadao NAKAI Tetsuo IGUCHI Masaharu NAKAZAWA Toshiyuki IIDA Shinichi TAMURA
We have developed a neutron imaging system based on the penumbral imaging technique. The system consists of a penumbral aperture and a sensitive neutron detector. The aperture was made from a thick (6 cm) tungsten block with a toroidal taper. It can effectively block 14-MeV neutrons and provide a satisfactory sharp, isoplanatic (space-invariant) point spread function (PSF). A two-dimensional scintillator array, which is coupled with a gated two-stage image intensifier system and a CCD camera, was used as a sensitive neutron detector. It can record the neutron image with high sensitivity and high signal-to-noise ratio. The reconstruction was performed with a Wiener filter. The spatial resolution of the reconstructed neutron image was estimated to be 31 µm by computer simulation. Experimental demonstration has been achieved by imaging 14-MeV deuterium-tritium neutrons emitted from a laser-imploded target.
This brief proposes a solar-cell-assisted wireless biosensing system that operates using a biofuel cell (BFC). To facilitate BFC area reduction for the use of this system in area-constrained continuous glucose monitoring contact lenses, an energy harvester combined with an on-chip solar cell is introduced as a dedicated power source for the transmitter. A dual-oscillator-based supply voltage monitor is employed to convert the BFC output into digital codes. From measurements of the test chip fabricated in 65-nm CMOS technology, the proposed system can achieve 99% BFC area reduction.
Chun-Liang LEE Chi-Wei CHEN Yaw-Chung CHEN
The differentiated services (Diffserv) architecture is a potential solution for providing quality of service (QoS) on the Internet. Most existing studies focus on providing service differentiation among few service classes. In this paper, we propose an approach which can achieve per-flow weighted fair rate allocation in a differentiated services network. Following the design philosophy of the Diffserv model, in the proposed approach core routers do not need to keep per-flow information. An edge router adjusts the transmission rate of a flow based on the feedback carried on control packets, which are inserted by the ingress edge router and returned by the egress edge router. Core routers periodically estimate the fair share rate of each virtual flow and mark the results in control packets. We use both simulations and analysis to evaluate the performance of the proposed approach. The analytical results show that our approach allows a system to converge to weighted fair rate allocations in limited time. Through the simulation results, we can further validate the analytical results, and demonstrate that better throughput can be achieved.
Wei CHEN Erry GUNAWAN Kah Chan TEH
Space-time array manifold model is usually used in a fast fading channel to estimate delay for the radio location. The existing additive white Gaussian noise (AWGN) estimation error model significantly overestimates the delay estimation. In this paper, we model the estimation error of the space-time array manifold channel impulse response (CIR) matrix as a correlated AWGN matrix and its performance is shown to be closer to the estimation error of practical systems than the existing model.
Guowei CHEN Xujiaming CHEN Kiichi NIITSU
This brief presents a slope analog-digital converter (ADC)-based supply voltage monitor (SVM) for biofuel-cell-powered supply-sensing systems operating in a supply voltage range of 0.18-0.35V. The proposed SVM is designed to utilize the output of energy harvester extracting power from biological reactions, realizing energy-autonomous sensor interfaces. A burst pulse generator uses a dynamic leakage suppression logic oscillator to generate a stable clock signal under the sub-threshold region for pulse counting. A slope-based voltage-to-time converter is employed to generate a pulse width proportional to the supply voltage with high linearity. The test chip of the proposed SVM is implemented in 180-nm CMOS technology with an active area of 0.018mm2. It consumes 2.1nW at 0.3V and achieves a conversion time of 117-673ms at 0.18-0.35V with a nonlinearity error of -5.5/+8.3mV, achieving an energy-efficient biosensing frontend.
Han-Yan WU Ling-Hwei CHEN Yu-Tai CHING
Embedding efficiency is an important issue in steganography methods. Matrix embedding (1, n, h) steganography was proposed by Crandall to achieve high embedding efficiency for palette images. This paper proposes a steganography method based on multilayer matrix embedding for palette images. First, a parity assignment is provided to increase the image quality. Then, a multilayer matrix embedding (k, 1, n, h) is presented to achieve high embedding efficiency and capacity. Without modifying the color palette, hk secret bits can be embedded into n pixels by changing at most k pixels. Under the same capacity, the embedding efficiency of the proposed method is compared with that of pixel-based steganography methods. The comparison indicates that the proposed method has higher embedding efficiency than pixel-based steganography methods. The experimental results also suggest that the proposed method provides higher image quality than some existing methods under the same embedding efficiency and capacity.
Wen SHI Jianling LIU Jingyu ZHANG Yuran MEN Hongwei CHEN Deke WANG Yang CAO
Syndrome is a crucial principle of Traditional Chinese Medicine. Formula classification is an effective approach to discover herb combinations for the clinical treatment of syndromes. In this study, a local search based firefly algorithm (LSFA) for parameter optimization and feature selection of support vector machines (SVMs) for formula classification is proposed. Parameters C and γ of SVMs are optimized by LSFA. Meanwhile, the effectiveness of herbs in formula classification is adopted as a feature. LSFA searches for well-performing subsets of features to maximize classification accuracy. In LSFA, a local search of fireflies is developed to improve FA. Simulations demonstrate that the proposed LSFA-SVM algorithm outperforms other classification algorithms on different datasets. Parameters C and γ and the features are optimized by LSFA to obtain better classification performance. The performance of FA is enhanced by the proposed local search mechanism.
Karunanithi BHARANITHARAN Jiun-Ren DING Bo-Wei CHEN Jhing-Fa WANG
In H.264/AVC intra frame coding, the rate-distortion optimization (RDO) is employed to select the optimal coding mode to achieve the minimum rate-distortion cost. Due to a large number of combinations of coding modes, the computational burden of Rate distortion optimization (RDO) becomes extremely high in intra prediction. In this paper, we proposed an efficient selective intra block size decision (SIB) that selects the appropriate block size for intra prediction, further proposed fast intra prediction algorithm reduces a number of modes required for RDO that significantly reduces the encoder complexity. Experimental results show that the proposed fast mode decision algorithm reduces the encoding time by up to 68% with negligible video quality degradation.
Wen-Cheng YEN Hung-Wei CHEN Yu-Tong LIN
In this era of System-On-a-Chip (SOC) technology, a designable initial state is required. Thus, embedding low voltage and low power Power-On-Reset (POR) circuit on the SOC chip is important for the portable device. This paper proposes a new POR circuit with process and temperature compensations. A band-gap reference is used in this circuit to reduce the effect of the temperature and process variations. With 200 mV hysteretic design provides robust noise immunity against voltage fluctuations on the power supply. The POR circuit has been designed, simulated, and implemented. A test chip has been fabricated by using 0.18 µm single-poly triple-metal CMOS logical process. Measurement results show the rise threshold voltage Vrr has only a 3% variation under the temperature range from -40 to 125. The power consumption is 39 mW at the 1.8 V power supply. The chip size of the POR is 62 mm280 mm. Thus, this POR circuit has a great potential to apply to a low power supply system.
Wei CHEN Gang LIU Jun GUO Shinichiro OMACHI Masako OMACHI Yujing GUO
In speech recognition, confidence annotation adopts a single confidence feature or a combination of different features for classification. These confidence features are always extracted from decoding information. However, it is proved that about 30% of knowledge of human speech understanding is mainly derived from high-level information. Thus, how to extract a high-level confidence feature statistically independent of decoding information is worth researching in speech recognition. In this paper, a novel confidence feature extraction algorithm based on latent topic similarity is proposed. Each word topic distribution and context topic distribution in one recognition result is firstly obtained using the latent Dirichlet allocation (LDA) topic model, and then, the proposed word confidence feature is extracted by determining the similarities between these two topic distributions. The experiments show that the proposed feature increases the number of information sources of confidence features with a good information complementary effect and can effectively improve the performance of confidence annotation combined with confidence features from decoding information.