Search

In this paper, the fabrication of the through-silicon via (TSV) is introduced and characterized. Pre-bond structural analysis and electrical tests were performed and analyzed to examine the robustness of the fabrication process and the structural integrity of the TSVs. It is shown that the TSVs are robust even after given stress. The influence of the presence of TSVs on nearby metal-oxide-semiconductor field-effect transistor (MOSFET) devices and metal interconnects is also characterized by means of electrical tests. The results show that the fabricated TSV is robust and that the presence of TSV does not pose significant impact on the performance of MOSFET as long as they are sufficiently distanced from each other.

Semantic segmentation has been one of the most important research areas in computer vision. In semantic segmentation, both the context information and spatial information are important in semantic segmentation performance of a con-volutional neural network (CNN) model. This study focuses on methods to optimize the context information. We proposed a structure consisting of multiple upsampling blocks and concatenated Dense Block(DB), enhancing the context information on the decoder by the fusion of multiple high-level features. Furthermore, we use Blur Pooling [1], which is a better downsampling method proposed by Richard Zhang, making the CNN more shift-invariant to an input image. Finally, we determined the drawback of using the cross entropy, which can result in poor small-object-detection performance, on semantic segmentation tasks. Furthermore, we boosted the performance of our model by applying the soft-IoU loss function. Finally, we test our model's performance using the CamVid dataset established. The experimental results show that our proposed approach can provide a better ability of integrating the context feature in the upsampling path. The mIoU result of our proposed model was 70.534% on the Camvid dataset, outperforming the 65.8% mIoU result of FC-DenseNet67, which is our baseline model, obtained from the study in [2].

With the rapid development and reduced cost of digital video capturing devices, driving video recorders (DVRs) begin to gain widespread popularity. What is seen and what happens along the way can thus be recorded in videos. However, searching for a specific scene or event among such massive video collections is laborious and tedious. In this paper, we develop a speed-adaptive street view image generation system using general front-mounted DVRs, requiring no additional devices deployed. Visual summaries of street scenes along the way can be provided, allowing users to retrieve a video clip corresponding to a specific road section quickly. An efficient algorithm for estimating the distance a pixel has moved between two consecutive frames is also proposed, so a street view image can be generated with an appropriate aspect ratio without demanding a constant driving speed. Experiments on an extensive data set show that our proposed system can efficiently generate street view images under different lighting and weather conditions, demonstrating its feasibility.

Since a proficient golf swing is a key element in success in golf, golfers always make much effort in improving their swing skills. However, if not under the supervision of a coach, a golfer cannot effectively perceive his/her faults in swings and comprehend how to change the swing mechanics. In this paper, we propose a vision-based golf training system, named "Improve My Golf Swing," which aims at improving the golfer's swing skills using intelligent glasses serving as a wearable displayer. The golfer's posture is analyzed through computation-efficient image processing techniques. Involving professional golf training knowledge, the proposed system automatically generates feature lines to evaluate whether the golfer's body is properly aligned. During a swing, the golfer can watch his/her own posture as well as the augmented feature lines displayed on the intelligent glasses. Instead of quantitative swing analysis, we focus on providing instructions of posture adjustment to the golfer in a more intuitive and unambiguous manner. Different audio alerts will be given to the golfer according to which part of a posture fault is detected. Experiments on different golf swing videos show encouraging results.

The traditional video annotation approaches focus on annotating keyframes, shots, or the whole video with semantic keywords. However, the extractions of keyframes and shots lack of semantic meanings, and it is hard to use a few keywords to describe a video by using multiple topics. Therefore, we propose a novel video annotation framework using near-duplicate segment detection not only to preserve but also to purify the semantic meanings of target annotation units. A hierarchical near-duplicate segment detection method is proposed to efficiently localize near-duplicate segments in frame-level. Videos containing near-duplicate segments are clustered and keyword distributions of clusters are analyzed. Finally, the keywords ranked according to keyword distribution scores are annotated onto the obtained annotation units. Comprehensive experiments demonstrate the effectiveness of the proposed video annotation framework and near-duplicate segment detection method.

In several applications, sequence databases generally update incrementally with time. Obviously, it is impractical and inefficient to re-mine sequential patterns from scratch every time a number of new sequences are added into the database. Some recent studies have focused on mining sequential patterns in an incremental manner; however, most of them only considered patterns extracted from time point-based data. In this paper, we proposed an efficient algorithm, Inc_TPMiner, to incrementally mine sequential patterns from interval-based data. We also employ some optimization techniques to reduce the search space effectively. The experimental results indicate that Inc_TPMiner is efficient in execution time and possesses scalability. Finally, we show the practicability of incremental mining of interval-based sequential patterns on real datasets.

With the explosive growth of the social multimedia sharing, copyright protection and search result refinement are always the critical issues for the service operators. To resolve the problems, content-based near-duplicate video retrieval is developed in recent years. In this paper, we construct a condensed Pattern-based Prefix tree (PP-tree) to index the patterns of reference videos for fast retrieval. To calculate how likely a query video and a reference video are near-duplicates, a novel algorithm for discovering the temporal relations among patterns is proposed. Comprehensive experiments on public datasets are conducted to verify the effectiveness and efficiency of the proposed method. Experimental results show that the proposed near-duplicate video retrieval approach outperforms the state-of-the-art approaches in terms of precision, recall, and execution time.

The equivalent oxide thickness in Ge MOS device is scaled down to 0.39 nm, and the leakage current is decreased as well. The improvement can be attributed to the in-situ Ge sub-oxide desorption process in an ALD chamber at 370 °C. About 95% Ge4+ in HfGeO x interfacial layer are obtained by H 2 O plasma process together with in-situ desorption before atomic layer deposition.

Through Silicon Interposer (TSI) needs to fulfill multi-die stacking in one packaging which can bring high integration density, short interconnection length and small size for next generation devices. Die stacking is a key process in the TSI manufacturing flow, and within that process, die warpage is of central concern. This is because the large warpage of the Si-mterposer induces poor joining of u-bump interconnection, lithograph missing and die breakage during the assembly process. According to our experience, the Cu metal density of redistribution layer (RDL) significantly effects on TSI warpage. [2] The work proposes to design four different Cu metal densities, namely code A (10~25%), B (25~40%), C (40~55%) and D (55~70%) and combine the different metal density to investigate the relationship between Cu metal pattern density and TSI warpage under four different TSI size. Each unit of Cu metal density area is 12 × 12 mm 2 . The TSI size is ranging from unit 1 × 1 (12 × 12 mm 2 ), 1 × 2 (12 × 24 mm 2 ), 2 × 2 (24 × 24 mm 2 ) and 2 × 3 (24 × 36 mm 2 ). In this work, the TSI warpage modeling methodology has been developed and carried out to find the correlation between pattern density and warpage behavior using the finite element method (FEM). A simplified pattern-inclusion model has been constructed to obtain its warpage response under a thermal loading. Furthermore, the equivalent model has been established to obtain the effective mechanical properties of the equivalent layer (i.e., the composite layer of Cu traces and IMD) through adapting the warpage results to correctly match the simplified pattern-inclusion model. Therefore, the TSI warpage can be calculated during the reflow process after acquiring the effective mechanical properties of different Cu pattern densities. The simulation warpage variation of 1 × 1 (12 × 12 mm 2 ) TSI die size for code A, B, C and D are 8 μm, 19 μm, 32 μm, and 47 μm, respectively. The simulation results indicate that large TSI with high Cu metal density possesses high warpage. After fabricating the different Cu metal density wafer, the wafer was grinded to 100 μm for dicing, then using thermal shadow moire to measure the TSI warpage after one time reflow process. The experimental TSI warpage variation of 1 × 1 (12 × 12 mm 2 ) TSI die size with code A, B, C and D are 6-10 μm, 21-26 μm, 31-36 μm, and 43-70 μm, respectively. The experimental results imply that large TSI with high Cu metal density possesses high warpage. After measuring the warpage of those dies during the thermal curve, we compared the measurement data with our simulation result and explored 3 subjects: different pattern density, die size, and symmetry. We found the TSI warpage between simulation and measurement result are consistent. The finite element simulation results in the work correlate well with the experimental test results. Therefore, we are highly confident that this finite element model can help generate design guidelines for wafer patterns and predict the die warpage variation during the assembly process.

This paper presents a speed controller which combines an integral controller with a Fuzzy controller to reduce the steady-state error of speed for a permanent magnet synchronous motor (PMSM). In this paper, the Space Vector PWM (SVPWM) principles are introduced, and modeling PMSM control system simulation on PSIM simulation software. Finally, the proposed method is implemented in a TMS320F28335 DSP-based speed Integral-Fuzzy control of PMSM, and the result has been proved that the proposed method was better than convention Fuzzy control.

In wiretap channels the eavesdropper's channel state information (CSI) is commonly assumed to be known at the transmitter, fully or partially. In practice, the eavesdropper may not be motivated to feedback correct CSI. In this paper we consider a more practical problem for the transmitter to have the eavesdropper's CSI. That is, the fast fading multiple-antenna Gaussian broadcast channels (FMGBC-CM) with confidential messages, where there are two receivers both legitimate and willing to feedback accurate CSI to the transmitter to maintain their secure transmission, and not to be eavesdropped by the other. We assume that only the statistics of the channel state information are known by the transmitter. We first derive the achievable rate region for the FMGBC-CM. After that we propose an iterative algorithm to solve the channel input covariance matrices and the inflation factor. Finally, numerical examples show that under perfect secrecy constraint both users can achieve positive rates simultaneously, which verifies our derived necessary condition. Numerical results also elucidate the effectiveness of the proposed algorithm in terms of the convergence speed.

Serve receive-to-attack (SR2A) is the most principal way to gain points in volleyball games. In addition, the positions of players on the court reveal informative clues about both offensive and defensive formations. Hence, in this paper, we propose an effective system capable of extracting SR2A periods from long broadcast volleyball videos (taken from a pan-tilt-zoom camera), and then locating the players using a novel histogram-based approach. Enriched visual presentation can be provided to give the audience or professional players/coaches a further sight into the game. The proposed system consists of four major processing modules, including court detection, SR2A period extraction, camera calibration, and histogram-based player localization. The experiments conducted on broadcast volleyball videos show promising results.

With the exponential growth of the web multimedia contents, illegal video copies are widespread and easy to be obtained from search engines and video sharing websites. For video copyright protection, near-duplicate video retrieval becomes more and more important. In this paper, we proposed a Pattern Set based Dynamic Programming (PSDP) algorithm to retrieve near-duplicate videos efficiently and effectively. In addition, the precise positions of the near-duplicate segments in videos can also be located. To better deal with the slow/fast motion and the dropped frame problem, the Time-shift Pattern set Similarity (TPS) is then applied. Two datasets are used to evaluate the effectiveness and efficiency of the proposed method, and the experimental results show that the proposed method outperforms the compared approaches in terms of precision and execution time in both two datasets.

Self-training plays an important role in sports exercise. However, if not under the instruction of a coach, improper training postures can cause serious harm to muscles and ligaments of the body. Hence, the development of computer-assisted self-training systems for sports exercise is a recently emerging research topic. In this paper, we propose a Yoga self-training system, entitled YogaST, which aims at instructing the user/practitioner to perform the asana (Yoga posture) correctly and preventing injury caused by improper postures. Involving professional Yoga training knowledge, YogaST analyzes the practitioner's posture from both front and side views using two Kinects with perpendicular viewing directions and assists him/her in rectifying bad postures. The contour, skeleton, and feature axes of the human body are extracted as posture representation. Then, YogaST analyzes the practitioner's posture and presents visualized instruction for posture rectification so that the practitioner can easily understand how to adjust his/her posture.

The purpose of this paper is to present a novel intelligent service robot for recognizes target object and transports it to the destination, which is called the “tRobot”. In order to achieve the tRobot's functions, we integrate the localization system, image recognition and tracking function, obstacle avoidance, and robotic arm into the tRobot. By the localization system, not only can locating the tRobot's current coordinate location, but also can provide information for places the object to a preset location that was set by user, when the object was be caught by the robot arm. The image recognition function can recognize and track target object. In terms of the obstacle avoidance and clamp function, we use four infrared (IR) sensors for distance measure and place them around the tRobot. When a short enough distance is sensed by IR sensor which is in the front of the tRobot, the clip on the robot arm will clamp the object and place it at the preset coordinate location. Finally, we integrate foregoing four functions into the tRobot, the results of the experimental shows available and good compatibility.

The TaN/HfON/GeO2/n-Ge pMOSFETs were fabricated with different formation processes of GeO2 interfacial layer. Ultra low EOT of around 0.5 nm is achieved using GeO2 grown by H2O plasma together with in-situ grown HfON gate dielectric, and simultaneously the peak hole mobility of Ge pMOSFET is 312 cm2/V*s.

In this paper we consider the secure transmission over the fast fading multiple antenna Gaussian broadcast channels with confidential messages (FMGBC-CM), where a multiple-antenna transmitter sends independent confidential messages to two users with information theoretic secrecy and only the statistics of the receivers' channel state information are known at the transmitter. We first use the same marginal property of the FMGBC-CM to classify the non-trivial cases, i.e., those not degraded to the common wiretap channels. We then derive the achievable rate region for the FMGBC-CM by solving the channel input covariance matrices and the inflation factor. Due to the complicated rate region formulae, we resort to low SNR analysis to investigate the characteristics of the channel. Finally, the numerical examples show that under the information-theoretic secrecy requirement both users can achieve positive rates simultaneously., Comment: 6 pages, conference, 5 figures

In the social network, living photos occupy a large portion of web contents. For sharing a photo with the people appearing in that, users have to manually tag the people with their names, and the social network system links the photo to the people immediately. However, tagging the photos manually is a time-consuming task while people take thousands of photos in their daily life. Therefore, more and more researchers put their eyes on how to recommend tags for a photo. In this paper, our goal is to recommend tags for a query photo with one tagged face. We fuse the results of face recognition and the useri¦s relationships obtained from social contexts. In addition, the Community-Based Group Associations, called CBGA, is proposed to discover the group associations among users through the community detection. Finally, the experimental evaluations show that the performance of photo tagging recommendation is improved by combining the face recognition and social relationship. Furthermore, the proposed framework achieves the high quality for social photo tagging recommendation.

In volleyball matches, spiking is the most effective way to gain points, while blocking is the action to prevent the opponents from getting scores by spiking. In this paper, we propose an intelligent system for automatic spiking events detection and blocking pattern classification in real volleyball videos. First, the entire videos are segmented into clips of rallies by whistle detection. Then, we find the court region based on proper camera calibration, and detect the location of the net for judging the positions of spiking and blocking. Via analyzing the changes of moving pixels along the net, we make a bounding box around the blocking location, so as to classify the blocking patterns into two main categories based on the width of bounding box. Finally, two important tactic patterns, delayed spiking and alternate position spiking, are recognized. With the information of spiking events and blocking locations, we can collect the statistical data and make tactics inference easily. To the best of our knowledge, no previous work is focused on spiking or blocking event detection. The experimental results on the videos recorded by a university volleyball team are promising and demonstrate the effectiveness of our proposed scheme.

Electrical characteristics and reliability of MOS devices with HfO 2 or Ti/HfO 2 /Hf high-k dielectrics are studied and compared in this work. For the MOS device with Ti/HfO 2 /Hf gate dielectric stack after a PDA at 600°C, its equivalent oxide thickness (EOT) value is 7.1 A and the leakage density is about 2.05 × 10−1A/cm2.

The HfO 2 based insulator is promising for resistive random access memory (RRAM). At first in this work, the effects of different oxygen contents in sputtered HfO 2 on resistive switching memory are investigated. It is observed that the oxygen flux rate would strongly affect the LRS/HRS current ratio and operation voltage. The switching behavior can be attributed to oxygen vacancy assisted conduction filament formation and annihilation. Then, the effects of Ti and Zr capping metal layers on RRAM devices with ALD-HfO 2 insulator are studied. Resistive switching behavior, temperature-dependence retention, and cycling endurance are strongly affected by the property of capping metal layer.

In basketball games, screen is a blocking move performed by an offensive player, who stands beside or behind a defender, in order to free a teammate to shoot, to receive a pass, or to drive in to score. Screen is the fundamental essence that most offensive tactics are executed with. In this paper, a screen-strategy analysis system is designed, and through combining the identified screens, what tactics are executed in basketball games can be speculated. The proposed system is capable of court region detection, camera calibration and player extraction. Player trajectories are computed by a Kalman filter-based tracking method and mapped to the real-world court coordinates. The player position/trajectory information greatly assists professional-oriented applications such as screen-strategy analysis and tactic inference. The experiments on broadcast basketball videos show encouraging results.

Video retrieval has been a hot topic due to the prevalence of video capturing devices and media-sharing services such as YouTube. Until now, few past studies has focused on querying the videos by images due to the semantic gap between images and videos is not easy to narrow. To this end, in this paper, we propose a novel semantic video retrieval system that integrates web image annotation and concept matching function to bridge images, concepts and videos. For web image annotation, we exploit textual and visual information in the web image to achieve effective image annotation. For concept matching function, we identify the concept relations by calculating the similarity between two concepts via Word Net. On the basis of web image annotation and concept matching function, the proposed system reaches the goals of usability and intelligence on semantic video retrieval. The experimental results reveal that our proposed system can successfully capture the user's intention between image concepts and video concepts for semantic video retrieval.

Free viewpoint video presentation is a new challenge in multimedia analysis. This paper presents an innovative physics-based scheme to reconstruct the 3D ball trajectory from single-camera volleyball video sequences for free viewpoint virtual replay. The problem of 2D-to-3D inference is arduous due to the loss of 3D information in projection to 2D images. The proposed scheme incorporates the domain knowledge of court specification and the physical characteristics of ball motion to accomplish the 2D-to-3D inference. Motion equations with the parameters are set up to define the 3D trajectories based on physical characteristics. Utilizing the geometric transformation of camera calibration, the 2D ball coordinates extracted over frames are used to approximate the parameters of the 3D motion equations, and finally the 3D ball trajectory can be reconstructed from single-camera sequences. The experiments show promising results. The reconstructed 3D trajectory enables the free viewpoint virtual replay and enriched visual presentation, making game watching a whole new experience.

In baseball games, different release points of pitchers form several kinds of pitching styles. Different pitching styles possess individual advantages. This paper presents a novel pitching style recognition approach for automatic generation of game information and video annotation. First, an effective object segmentation algorithm is designed to compute the body contour and extract the pitcher's body. Then, star skeleton is used as the representative descriptor of the pitcher posture for pitching style recognition. The proposed approach has been tested on broadcast baseball video and the promising experimental results validate the robustness and practicability.

Recently, more and more organizations attempt to provide service businesses to increase their customer satisfaction, organizational profit, and market share. Despite service businesses have contributed to the economic growth and employment in many countries, their productivity growth is still stagnating. In the resource-based view (RBV), the resources and capabilities of an organization are seen as the leverage for developing the core competencies, which are the key to create and sustain its competitive advantage. To quote an old sage: “No measurement, no management.” Measuring the core competencies of service businesses is vital to the competitive advantages of organizations around the world. Based on the RBV, this paper aims at proposing a model to measure and improve the core competencies of service businesses, with various indexes of the resources and capabilities. The model encompasses the recently proposed Service Innovation Capability Maturity Model (SICMMTM) and will be illustrated through case studies. Finally, recommendation for practice will be presented and discussed.

We propose a new design of 25-GHz flat-top multichannel optical filter. The flat-top of optical filters can improve to 2.288×10−11dB which is superior to that of 5 orders of previous study. The superior performance of dispersion and channel spacing isolation improve 73.5% and 94%, respectively, which compared with previous study.

A single inductor matching network that carried low noise is designed to achieve the input wideband matching. This way has lower complexity that reduces chip area and holds the good reflection coefficient. Besides, the current reuse technique was used to achieve low power consumption. The design is simulated by Taiwan Semiconductor Manufacturing Company (TSMC) 0.18µm RF CMOS process. Through a 1V/5.56mA supply, The measurement results show that the LNA achieved the maximum gain of 14.5dB with gain flatness ± 0.35dB; input return loss lower than −10dB; and a minimum noise figure 2.9dB in 3∼5 GHz.

We describe the educational cloud layer of our Cadabia middleware. This layer will permit users to specify the machines which will be used for copies of the materials for various groups that they have joined. This kind of “personal cloud” will help to protect data for both individuals and institutions, while making other data searchable by those who have used similar ontologies.

Up to the present, the contemporary studies are not really successful in image annotation due to some critical problems like diverse regularities between visual features and human concepts. Such diverse regularities make it hard to annotate the image semantics correctly. In this paper, we propose a novel approach called AICDM (Annotation by Image-Concept Distribution Model) for image annotation by discovering the associations between visual features and human concepts from image-concept distribution. Through the proposed image-concept distribution model, the uncertain regularities between visual features and human concepts can be clarified for achieving high-quality image annotation. The empirical evaluation results also reveal that our proposed AICDM method can effectively alleviate the uncertain regularity problem and bring out better annotation results than other existing approaches in terms of precision and recall.

A single inductor matching network that carried low noise is designed to achieve the input wideband matching. This way has lower complexity that reduces chip area and holds the good reflection coefficient. Besides, the current reuse technique was used to achieve low power consumption. The design is simulated by Taiwan Semiconductor Manufacturing Company (TSMC) 0.18µm RF CMOS process. Through a 1V/5.56mA supply, the LNA achieved the maximum gain of 16.69dB with gain flatness ± 0.5dB; input return loss lower than −10dB; and a minimum noise figure 2.6dB in 3∼5 GHz.

Electrical characteristics and thermal stability of MOS devices with TiN/MoN metal stacks and various gate dielectrics were studied in this work. High work function gate is achieved by integrating TiN/MoN stack with HfAlO dielectric. Reliability characteristics, in terms of stress-induced leakage current and stress-induced Vfb shift, are improved for metal gate stack on high-k dielectric in comparison with that on SiO 2 .

In this study, a new piezomotor is presented, and an optical anti-shake image system using this piezomotor is also developed. Comparing with other linear piezomotors, this motor is simpler, compacter, and more powerful. The dimension of the motor is 16mm(L)×8mm(W)×1.5mm(H), and the resonate frequency is 220kHz. Moreover, the push force at operating voltage 10Vpp is 15g, and the power consumption is 0.23W. Because of the long travel range:±2mm and high maximum velocity:26mm/s, it could satisfy the compensative displacement of the CMOS sensor caused by the hand tremor in a anti-shake image system. In this paper, a dynamic behavior of the piezomotor is investigated by COMSOL software, and the control algorithm in the optical anti-shake image system, including tremor sensing, motor position sensing, is also described.

As generally recognized, for the conventional power-factor-correction (PFC) boost converter operating in the continuous current mode (CCM), most of power dissipation comes from the reverse recovery current of the boost diode. Consequently, a simple zero-current-switching (ZCS) circuit, based on a passive circuit without any complicated switch control, is presented herein and applied to the boost diode of the PFC boost converter operating in CCM, so as to increase the total efficiency as well as to reduce the electromagnetic interference (EMI). Some simulation and experimental results are provided to verify the effectiveness of the proposed soft-switching topology.

Contact resistance (Rc) between metal and carbon nanotubes (CNTs) is studied extensively. Metal oxide formation at interface due to oxygen absorption plays very important role. Chemically inert metals such as Pt and Au results in the lowest Rc. Adding Ta into Pt can solve the adhesion issue while keeps low Rc . Those metals can break metal oxide and form metal carbide are also preferred

A sigma-delta modulation controlled DC/AC inverter is proposed. Modeling and analysis of sigma-delta modulation (SDM) controller are explored. The pulse width of the SDM formed by cycle-by-cycle integration loop is clearly described. System stability is studied by root-locus and Bode plot techniques. An approach to reducing the THD is conducted by Fourier analysis. A prototype of 1 kVA SDM controlled DC/AC inverter is designed and examined. Simulation and experiment are close to the theoretical predictions. The proposed inverter can provide high efficiency up to 90% under the rated power and still keep the total harmonic distortion (THD) without EMI filter within 2%.

A multi-purpose real-time visual tracking system that can be used in different scenarios has been designed and implemented in this study. This real-time visual tracking system can be operated in two modes - "static tracking" and "dynamic tracking". When operated in the static tracking mode, a modified adaptive background subtraction method is used to detect the moving object. In contrast, when operated in the dynamic tracking mode, a region-based matching approach is used to estimate the motion of the moving target and the servo control unit would be activated to control the camera motion so that the moving target's image is kept in the center of the monitoring screen. Several experiments such as "traffic monitoring", "marine creature detection", and "intruder detection/tracking" were conducted to evaluate the system's performance. Experimental results indicate that the performance of the developed real-time visual tracking system is satisfactory.