A number of well-known learning-based face detectors can achieve extraordinary performance in controlled environments. But face detection under varying illumination is still challenging. Possible solutions to this illumination problem could be creating illumination invariant features or utilizing skin color information. However, the features and skin colors are not sufficiently reliable under difficult lighting conditions. Another possible solution is to do illumination normalization (e.g., Histogram Equalization (HE)) prior to executing face detectors. However, applications of normalization to face detection have not been widely studied in the literature. This paper applies and evaluates various existing normalization methods under the framework of combining the illumination normalization and two learning-based face detectors (Haar-like face detector and LBP face detector). These methods were initially proposed for different purposes (face recognition or image quality enhancement), but some of them significantly improve the original face detectors and lead to better performance than HE according to the results of the comparative experiments on two databases. Meanwhile, we propose a new normalization method called segmentation-based half histogram stretching and truncation (SH) for face detection under varying illumination. It first employs Otsu method to segment the histogram (intensities) of the input image into several spans and then does the redistribution on the segmented spans. In this way, the non-uniform illumination can be efficiently compensated and local facial structures can be appropriately enhanced. Our method obtains good performance according to the experiments.

The research reported in this paper is an attempt to elucidate the predictors of pause duration in read-aloud discourse. Through simple linear regression analysis and stepwise multiple linear regression, we examined how different factors (namely, syntactic structure, discourse hierarchy, topic structure, preboundary length, and postboundary length) influenced pause duration both separately and jointly. Results from simple regression analysis showed that discourse hierarchy, syntactic structure, topic structure, and postboundary length had significant impacts on boundary pause duration. However, when these factors were tested in a stepwise regression analysis, only discourse hierarchy, syntactic structure, and postboundary length were found to have significant impacts on boundary pause duration. The regression model that best predicted boundary pause duration in discourse context was the one that first included syntactic structure, and then included discourse hierarchy and postboundary length. This model could account for about 80% of the variance of pause duration. Tests of mediation models showed that the effects of topic structure and discourse hierarchy were significantly mediated by syntactic structure, which was most closely correlated with pause duration. These results support an integrated model combining the influence of several factors and can be applied to text-to-speech systems.

This paper proposes a multipacket-per-slot reservation-based random access protocol with multiuser detection (MD) and automatic repeat request (ARQ), called MPRMD, and analyzes its performance by computer simulations. In MPRMD, before data packet (DP) transmission, a user terminal (UT) transmits a small access request packet (AP) that is composed of an orthogonal preamble sequence and a UT identifier (UT-ID) in a randomly selected minislot during a short dedicated period. Even when several APs collide, a base station (BS) distinguishes them by matched filtering against the preamble part and then extracts the UT-IDs after separating each AP by MD. If the APs are not successfully detected, a small number of minislots are additionally arranged to retransmit them. Thus, by using MD under AP crowded conditions, BS can maximally detect the access requests in a short period, which results in reducing the overhead. Furthermore, in the assignment of a slot, BS intentionally assigns one slot to multiple UTs in order to enhance the efficiency and separates UT's DPs by MD. Since MPRMD can detect a multitude of access requests by utilizing MD in the short period and efficiently assign the slot to separable DPs by MD, it can enhance the system throughput. Computer simulations are conducted to demonstrate the effectiveness of MPRMD. It is shown that the maximum throughputs of MPRMD with the average SNR of 30dB reach 1.4 and 1.7 packets/slot when a data packet is 10 times and 50 times as long as a control packet, respectively.

Due to the recent technical advances, GPUs are used for general applications as well as screen display. Many research results have been proposed to the performance of previous CPU-based algorithms by a few hundred times using the GPUs. In this paper, we propose a density-based clustering algorithm called GSCAN, which reduces the number of unnecessary distance computations using a grid structure. As a result of our experiments, GSCAN outperformed CUDA-DClust [2] and DBSCAN [3] by up to 13.9 and 32.6 times, respectively.

This letter presents a novel phase synchronization algorithm for a MIMO radar system in order to overcome the limitation of the existing algorithms relying on channel reciprocity, or line-of-sight, assumption between radar elements. The proposed algorithm is capable of synchronizing local oscillator phases among radar elements even if line-of-sight communication links are not available. Furthermore, the proposed algorithm exhibits robust MSE performance in the presence of frequency estimation error. The performance of the proposed algorithm was analyzed theoretically and verified by simulations.

Excimer laser annealing at 308nm in UV and semiconductor blue laser-diode annealing at 445nm were performed and compared in term of the crystallization depending on electrical properties of Si films. As a result for the thin Si films of 50nm thickness, both lasers are very effective to enlarge the grain size and to activate electrically the dopant atoms in the CVD Si film. Smooth Si surface can be obtained using blue-laser annealing of scanned CW mode. By improving the film quality of amorphous Si deposited by sputtering for subsequent crystallization, both laser annealing techniques are effective for LTPS applications not only on conventional glass but also on flexible sheet. By conducting the latter advanced annealing technique, small grain size as well as large grains can be controlled. As blue laser is effective to crystallize even rather thicker Si films of 1µm, high performance thin-film photo-sensor or photo-voltaic applications are also expected.

For the past couple of decades, the usage of the Web as a platform for deploying software products has become incredibly popular. Web applications became more prevalent, as well as more complex. Countless Web applications have already been designed, developed, tested, and deployed on the Internet. However, it is noticeable that many common functionalities are present among these vast number of applications. This paper proposes an approach based on a database containing information from previous test artifacts. The information is used to generate test scenarios for Web applications under test. We have developed a tool based on our proposed approach, with the aim of reducing the effort required from software test engineers and professionals during the test planning and creation stage of software engineering. We evaluated our approach from three viewpoints: comparison between our approach and manual generation, qualitative evaluation by professional software engineers, and comparison between our approach and two open-source tools.

A suffix tree is widely adopted for indexing genome sequences. While supporting highly efficient search, the suffix tree has a few shortcomings such as very large size and very long construction time. In this paper, we propose a very fast parallel algorithm to construct a disk-based suffix tree for human genome sequences. Our algorithm constructs a suffix array for part of the suffixes in the human genome sequence and then converts it into a suffix tree very quickly. It outperformed the previous algorithms by Loh et al. and Barsky et al. by up to 2.09 and 3.04 times, respectively.

A 125MHz 64-phase delay-locked loop (DLL) is implemented for time recovery in a digital wire-line system. The architecture of the proposed DLL comprises a coarse-locking circuit added to a conventional DLL circuit, which consists of a delay line including a bias circuit, phase detector, charge pump, and loop filter. The proposed coarse-locking circuit reduces the locking time of the DLL and prevents harmonic locking, regardless of the duty cycle of the clock. In order to verify the performance of the proposed coarse-locking circuit, a 64-phase DLL with an operating frequency range of 40 to 200MHz is fabricated using a 0.18-µm 1-poly 6-metal CMOS process with a 1.8V supply. The measured rms and peak-to-peak jitter of the output clock are 3.07ps and 21.1ps, respectively. The DNL and INL of the 64-phase output clock are measured to be -0.338/+0.164 LSB and -0.464/+0.171 LSB, respectively, at an operating frequency of 125MHz. The area and power consumption of the implemented DLL are 0.3mm2 and 12.7mW, respectively.

This paper presents a G-band triple-push voltage controlled oscillator (VCO) operating around 177GHz. The VCO, implemented in a commercial 0.13-µm RFCMOS technology, adopts a triple-push topology that is composed of 3 symmetrically coupled identical Colpitts sub-oscillators. Oscillation frequency can be tuned from 175.9GHz to 178.4GHz with varactor tuning voltage swept from 0 to 1.2V. The calibrated output power ranged from -19.7dBm to -16.6dBm depending on the oscillation frequency. The measured phase noise of the VCO is -80.2dBc/Hz at 1MHz offset. The results clearly demonstrate the possibility of applying triple-push topology for VCOs operating beyond 100GHz, enabling various high frequency applications that require tunable frequency sources.

To better support data-intensive workflows which are typically built out of various independently developed executables, this paper proposes extensions to parallel database systems called User-Defined eXecutables (UDX) and collective queries. UDX facilitates the description of workflows by enabling seamless integrations of external executables into SQL statements without any efforts to write programs confirming to strict specifications of databases. A collective query is an SQL query whose results are distributed to multiple clients and then processed by them in parallel, using arbitrary UDX. It provides efficient parallelization of executables through the data transfer optimization algorithms that distribute query results to multiple clients, taking both communication cost and computational loads into account. We implement this concept in a system called ParaLite, a parallel database system based on a popular lightweight database SQLite. Our experiments show that ParaLite has several times higher performance over Hive for typical SQL tasks and has 10x speedup compared to a commercial DBMS for executables. In addition, this paper studies a real-world text processing workflow and builds it on top of ParaLite, Hadoop, Hive and general files. Our experiences indicate that ParaLite outperforms other systems in both productivity and performance for the workflow.

In vehicular ad hoc networks, the efficient and reliable dissemination of emergency messages in a highly mobile environment under dense or sparse network is a significant challenge. This paper proposes a new vehicular broadcast protocol, called ACK-CAST, that can operate effectively in both dense and sparse network scenarios. ACK-CAST relies on acknowledgment messages from one-hop neighbors to select the next rebroadcasting vehicle. Simulation results show that ACK-CAST outperforms the SERVUS protocol in terms of the end-to-end delay, message delivery ratio and network overhead.

An efficient beamforming algorithm for large-scale phased arrays with lossy digital phase shifters is presented. This problem, which arises in microwave power transmission from solar power satellites, is to maximize the array gain in a desired direction with the gain loss of the phase shifters taken into account. In this paper the problem is first formulated as a discrete optimization problem, which is then decomposed into element-wise subproblems by the real rotation theorem. Based on this approach, a polynomial-time algorithm to solve the problem numerically is constructed and its effectiveness is verified by numerical simulations.

As data volumes explode, data storage costs become a large fraction of total IT costs. We can reduce the costs substantially by using compression. However, it is generally known that database compression is not suitable for write-intensive workloads. In this paper, we provide a comprehensive solution to improve the performance of compressed databases for write-intensive OLTP workloads. We find that storing data too densely in compressed pages incurs many future page splits, which require exclusive locks. In order to avoid lock contention, we reduce page splits by sacrificing a couple of percent of space savings. We reserve enough space in each compressed page for future updates of records and prevent page merges that are prone to incur page splits in the near future. The experimental results using TPC-C benchmark and MySQL/InnoDB show that our method gives 1.5 times higher throughput with 33% space savings compared with the uncompressed counterpart and 1.8 times higher throughput with only 1% more space compared with the state-of-the-art compression method developed by Facebook.

This paper gives a write voltage and read reference current generator considering temperature characteristics for multi-level Ge2Sb2Te5-based phase change memories. Since the optimum SET and RESET voltages linearly changes by the temperature, the voltage supply circuit must track this characteristic. In addition, the measurement results show that the read current depends on both read temperature and the write temperature and has exponential dependence on the read temperature. Thus, the binning technique is applied for each read and write temperature regions. The proposed variable TC generator can achieve below ±0.5 LSB precision from the measured differential non-linearity and integral non-linearity. As a result, the temperature characteristics of both the linear write voltage and the exponential read current can be tracked with the proposed variation tolerant linear temperature coefficient current generator.

Probabilistic frequent itemset mining, which discovers frequent itemsets from uncertain data, has attracted much attention due to inherent uncertainty in the real world. Many algorithms have been proposed to tackle this problem, but their performance is not satisfactory because handling uncertainty incurs high processing cost. To accelerate such computation, we utilize GPUs (Graphics Processing Units). Our previous work accelerated an existing algorithm with a single GPU. In this paper, we extend the work to employ multiple GPUs. Proposed methods minimize the amount of data that need to be communicated among GPUs, and achieve load balancing as well. Based on the methods, we also present algorithms on a GPU cluster. Experiments show that the single-node methods realize near-linear speedups, and the methods on a GPU cluster of eight nodes achieve up to a 7.1 times speedup.

This paper proposes a radiation-hardened phase-locked loop (RH-PLL) with a switchable dual modular redundancy (DMR) structure. After radiation strikes, unhardened PLLs suffer clock perturbations. Conventional RH-PLLs have been proposed to reduce recovery time after perturbation. However, this recovery still requires tens of clock cycles. Our proposal involves ‘detecting’ and ‘switching’, rather than ‘recovering’ from clock perturbation. Detection speed is crucial for robust perturbation-immunity. We identify types of clock perturbation and then propose a set of detectors to detect each type. With this method, the detectors guarantee high-speed detection that leads to perturbation-immune switching from a radiated clock to an undistorted clock. The proposed RH-PLL was fabricated and then verified with a radiation test on real silicon.

In sentiment classification, conventional supervised approaches heavily rely on a large amount of linguistic resources, which are costly to obtain for under-resourced languages. To overcome this scarce resource problem, there exist several methods that exploit graph-based semi-supervised learning (SSL). However, fundamental issues such as controlling label propagation, choosing the initial seeds, selecting edges have barely been studied. Our evaluation on three real datasets demonstrates that manipulating the label propagating behavior and choosing labeled seeds appropriately play a critical role in adopting graph-based SSL approaches for this task.

In this paper, we propose an optimum access method for a phase change memory (PCM) with NAND strings. A PCM with a block erase interface is proposed. The method, which has a SET block erase operation and fast RESET programming, is proposed since the SET operation causes a slow access time for conventional PCM;. From the results of measurement, the SET-ERASE operation is successfully completed while the RESET-ERASE operation is incomplete owing to serial connection. As a result, the block erase interface with the SET-ERASE and RESET program method realizes a 7.7 times faster write speed compared than a conventional RAM interface owing to the long SET time. We also give pass-transistor design guidelines for PCM with NAND strings. In addition, the write-capability and write-disturb problems are investigated. The ERASE operation for the proposed device structure can be realized with the same current as that for the SET operation of a single cell. For the pass transistor, about 4.4 times larger on-current is needed to carry out the RESET operation and to avoid the write-disturb problem than the minimum RESET current of a single cell. In this paper, the SET programming method is also verified for a conventional RAM interface. The experimental results show that the write-capability and write-disturb problems are negligible.

We review our recent achievements in monolithic 3D-ICs and flexible electronics based on single-grain Si TFTs that are fabricated inside a single-grain with a low-temperature process. Based on pulsed-laser crystallization and submicron sized cavities made in the substrate, amorphous-Si precursor film was converted into poly-Si having grains that are formed on predetermined positions. Using the method called µ-Czochralski process and LPCVD a-Si precursor film, two layers of the SG Si TFT layers with the grains having a diameter of 6µm were vertically stacked with a maximum process temperature of 550°C. Mobility for electrons and holes were 600cm2/Vs and 200cm2/Vs, respectively. As a demonstration of monolithic 3D-ICs, the two SG-TFT layers were successfully implemented into CMOS inverter, 3D 6T-SRAM and single-grain lateral PIN photo-diode with in-pixel amplifier. The SG Si TFTs were applied to flexible electronics. In this case, the a-Si precursor was prepared by doctor-blade coating of liquid-Si based on pure cyclopentasilane (CPS) on a polyimide (PI) substrate with maximum process temperature of 350°C. The µ-Czochralski process provided location-controlled Si grains with a diameter of 3µm and mobilities of 460 and 121cm2/Vs for electrons and holes, respectively, were obtained. The devices on PI were transferred to a plastic foil which can operate with a bending diameter of 6mm. Those results indicate that the SG TFTs are attractive for their use in both monolithic 3D-ICs and flexible electronics.