Search

In this paper, we compare conventional saddle type FinFETs to partial isolation type saddle FinFETs (Pi-FinFETs) using 3D TCAD simulations to examine the effect of single charge traps for proper prediction of leakage current. We simulated single charge traps at various locations in the drain region, and analyzed how the traps affect leakage current. Our results show that Pi-FinFETs enhanced the leakage current characteristics given the presence of a single charge trap. Also, it was found that Pi-FinFETs exhibit half the $\text{F}_{\mathrm {TAT}}$ of S-FinFETs. Based on the results from our analysis method, where we use $\text{I}_{\mathrm {off}}$ fluctuation, the $\text{F}_{\mathrm {TAT}}$ , the $\sigma _{\mathrm {F}}$ and the $\text{P}_{\mathrm {F}}$ parameters to accurately compare performance, and present device design guidelines aimed at improving DRAM refresh characteristics.

This paper proposes A Design of charge pump for fast rising and falling time and adequate for Sigma Delta Modulation (SDM). Low Noise Phase-Locked Loop (PLL) architecture using reference clock quadrature is shown. The degree of noise performance improvement that can be obtained using clock quadrature is expressed in an expression. To fulfill the speed of the quadrupled reference clock, fully differential charge pump (CP) architecture is used. Also, Implemented the design of a unity gain buffer in charge pump for the optimized operation for Sigma Delta Modulation (SDM) which is also adequate for wide bandwidth PLL because of the high bandwidth and the ability to handle high load current. The proposed structure is implemented using CMOS 40nm process and uses 1.1V supply power.

Ambient backscatter communication (AmBC) is a promising solution for Internet of Things (IoT) network in terms of installation cost and energy efficiency. However, it is challenging to achieve massive connectivity in typical AmBC system due to lack of coordination among RF sources, tags and the reader. Such uncoordinated interferences can seriously degrade the system performance, limiting the data rate and communication range essential for numerous IoT applications. To tackle the challenging issues, we propose a novel sparse-coded AmBC (SC-AmBC) system using sparse codes and utilizing the structure of orthogonal frequency division multiplexing (OFDM) signals. The system utilizes the inherent characteristics in ambient sources and RF tags to enhance connectivity for massive IoT applications including smart city and factory. Numerical results demonstrate the potential of sparse codes for AmBC in terms of energy efficiency and receiver sensitivity at the reader.

Dialog state tracking in spoken dialog system is the task that tracks the flow of a dialog and grasps what a user wants from the utterance precisely. Since the dialog success is related to catching the want of the user, dialog state tracking is a necessary component for spoken dialog systems. This paper proposes a neural dialog state tracker with the attention mechanism for focusing on valuable words and the hierarchical softmax for efficient training of the tracker. In addition, the proposed tracker combines a natural language understanding module and a dialog state module in an end-to-end style. As a result, the error propagation within a dialog system is minimized. To prove the effectiveness of the proposed model, we do experiments on dialog state tracking in the human-human task-oriented dialogs. Our experimental results show that the proposed method outperforms both the neural tracker without the attention mechanism and that without the hierarchical softmax.

In this paper, we proposed novel test structures to evaluate the distribution of interface state density of MOSFETs by using optical charge pumping method. Unlike other measurement methods to extract interface state density (Dit), which have a limited range of measurable energy states and influenced by gate area and gate leakage, Dit can be extracted without these problems by using the proposed test structures. Test structures were fabricated using a 0.18μ CMOS process or FD-SOI technology with high-k dielectric, respectively. Optical charge pumping was performed in proposed test structures and Dit is extracted from 109 cm−2· eV−1 to 1013 cm−2· eV−1.

In this paper, Low Frequency Noise(LFN) characteristics of Fully Depleted Silicon On Insulator (FD-SOI) Tunneling Field Effect Transistor(TFET) are analyzed and the level of LFN in TFET is also compared with that of Metal Oxide Semiconductor Field Effect Transistor(MOSFET) fabricated with the same process. The level of LFN in TFET is observed much higher than that of MOSFET. That is because LFN in TFET is largely affected by the tunneling probability and fluctuation of tunneling probability is much more vulnerable to electric field induced traps than that of carrier number and mobility described as LFN mechanism of MOSFET.

A new biased proportional navigation guidance law considering the missile's dynamic lag is proposed to control the impact angle. The homing guidance loop having the first-order dynamic lag is modeled as the confluent hyper-geometric differential equation and its analytic solution is derived. In consequence, the guidance law reduces to finding an appropriate biased acceleration command which satisfies the given terminal constraints. A systematic way to determine the time-varying biased terms is devised based on the closed-form homing trajectory. In comparison with the existing homing guidance laws, the proposed method could improve the impact angle control performance.

Ambient backscatter is a promising solution for future Internet of Things (IoT) composed of massive low-power wireless devices without external power sources. It is challenging to accommodate these devices with high throughput and long range because of the low efficiency inherent in ambient backscatter. To tackle the challenging issue, we propose an optimum modulation and coding scheme (MCS) for ambient backscatter communication networks (ABCN). We model and analyze the ABCN embedded by semi-passive tags with M-ary modulation and full-duplex WiFi access points as reader. The coverage and network capacity are derived, by which the effect of various system parameters including reflection coefficient, code rate, and cluster size is investigated. It is shown that the optimum MCS for the ABCN will provide highly increased throughput and range for future IoT networks and smart home/office.

This paper is an approach for pedestrian detection and tracking with infrared imagery. The detection phase is performed by AdaBoost algorithm based on Haar-like features. AdaBoost classifier is trained with datasets generated from infrared images. The number of negative images used for training with AdaBoost algorithm is 3000. For positive training, 1000 samples are used After detecting the pedestrian with AdaBoost classifier, we proposed the Tracking-Learning-Detection (TLD) frameworks tracking strategies. TLD frameworks are preferred in this study because of its high accuracy rate and computation speed Tracking performance comparison is made between TLD and particle filtering. Results prove that TLD performs a higher tracking rate than particle filtering.

In this paper, we implemented the Side Scan Sonar (SSS) module for the UnderWater Simulator (UWSim) [1], specifically focusing on sonar image simulation. To obtain the simulated SSS images, we adopted the Lambertian diffusion model [2]. The beam profile was modeled using a pair of sine wave functions. We developed this sonar module in the UWSim which runs on Robot Operating System (ROS) framework. For validation, we applied feature matching to the obtained sonar images, and examined the inliers and matching performance.

Re-ranking models of parse trees have been focused on re-ordering parse trees with a syntactic view. However, also a semantic view should be considered in re-ranking parse trees, because the fact that a word pair has a dependency implies that the pair has both syntactic and semantic relations. This paper proposes a re-ranking model for dependency parsing based on a combination of syntactic and semantic plausibilities of dependencies. The syntactic probability is used as a syntactic plausibility of a parse tree, and a knowledge graph embedding is adopted to represent its semantic plausibility. The knowledge graph embedding allows the semantic plausibility of parse trees to be expressed effectively with ease. The experiments on the standard Penn Treebank corpus prove that the proposed model improves the base parser regardless of the number of candidate parse trees.

We demonstrate new design of Deep-UV LEDs extracting strong sidewall-emission redirect top direction. We observe considerably enhanced optical and improved electrical properties and expect this model can provide key idea in current DUV LEDs for enhancing light extraction efficiency.

The efficiency of an AlGaN deep-ultraviolet light-emitting diode with peak emission wavelength of 285nm is investigated as a function of current over a wide range of temperatures (110K to 300K). We find that the efficiency-versus-current curve exhibits unique and distinct features over the entire temperature range including three points of inflection, a u-shape phenomenon which the efficiency increases again after the minimum, and higher low-temperature efficiency than room-temperature efficiency at high-current density regime.

CZTSSe thin-films were deposited by stacked sputtering methods (ZnS/SnS/Cu) and annealed with selenization. We adjusted the thickness of the ZnS precursor layer in CZT precursors. A 337 nm thickness of ZnS precursor was shown an efficiency of up to 9.1%. We investigated the secondary phases by Raman spectroscopy and Kelvin probe force microscopy with depth profiles. The Cu2SnSe3, ZnSe, and MoSe2 secondary phases appeared near the back contact region. The phase distributions of the CZTSSe thin-films are different depending on ZnS precursor thickness with different depths. This phase characterization can describe the influences to the device performance of the CZTSSe thin-film solar cells.

The operating characteristics was calculated through a finite element analysis (Maxwell). Iron loss and copper loss was estimated through the calculation and the mechanical loss is the result of the experiment for the calculation of the efficiency. In this paper, the shape design of IPMSM(Interior Permanent Magnet Synchronous Motor) for improving the efficiency was carried out. As a result, Back EMF of the improved model is increased. So, copper loss was reduced comparing to the same output. Through the shape design for avoidance of the magnetic saturation, iron loss was reduced to improve efficiency.

Generally, the motor inside vehicle is exposed to highly ambient temperature and large vibration according to repeatedly starting and stopping for very short time. So, in this paper, the rotor shape design was performed to improve demagnetization endurance by considering the starting current of the Brushless DC (BLDC) Motor through the Finite Element Method(FEM). As a result, the partial irreversible demagnetization was occurred by starting current at the end of Permanent Magnets of the basic model. To solve this problem, the flux barrier design was performed and the flux on the rotor core was limited. Accordingly, demagnetization endurance and operating characteristics were improved.

In this study, we propose a new complementary metal-oxide semiconductor (CMOS) image sensor to extend the dynamic range. The proposed active pixel sensor (APS) uses a gate/body-tied PMOSFET-type (GBT) photodetector and an output voltage feedback structure. Although the size of the proposed APS is slightly larger than a conventional APS, the dynamic range of the proposed pixel is much wider than a conventional APS. And the proposed APS has a varied dynamic range from adjusting the reference voltage. The proposed APS has been designed and fabricated by using 0.35 μm 2-poly 4-metal standard CMOS technology, and its characteristics have been evaluated.

Since Charles Wheatstone introduced the stereoscope concept in 1838, 3D technology has encountered the 3rd golden age. Especially, after “Avatar” interests in 3D began to rapidly increase and now the 3D technology is widespread across a range of communication and entertainment platforms. However, the current 3D market has the limited 3D applications delivering the limited contents to passive consumers. To obtain the driving force in the 3D market, there will be many new challenges with respect to 3D TV ecosystem which consists of content creation, compression, processing, display and interaction. In this paper, several new technical challenges for future 3D based on 3D TV ecosystem will be defined and key enabling technologies will be suggested to solve the current issues. And finally new opportunities in the future 3D will be discussed.

Cu2ZnSn(S, Se)4 (CZTSSe) is emerged as a promising material because of non-toxic, inexpensive and earth abundant more than Cu(In, Ga)Se2. The highest conversion efficiency of CZTSSe is 11.1 % at IBM based on hydrazine process [1]. We have achieved CZTSSe thin-film solar cell 8% conversion efficiency by sputtering process. CZTSSe thin-films show different efficiency depending on selenization process. Depending on the different selenization process, conversion efficiencies in each sample are definitely distinguished from 8.06 to 3.17%. We investigated local electrical properties on these samples. The local surface potential is also critically different, which is ~ 200 mV on 8.06% film and ~ 70 mV on 3.17% film. From these results, we can suggest that selenization process can affect to local electrical characteristic as well as improving solar cell performances.

The paper proposes a novel recognition system for hand shapes at a distance for smart TV. Two types of hand shapes are selected for the needs of TV browsing. The proposed method provides robust recognition performance on various hand orientations and guarantees real-time computation.

In this paper, we propose a fast motion deblurring technique based on a motion blur region search for a mobile camera phone. The technique simultaneously captures two differently exposed images in one shot. By comparing edge information from short and normal exposed images, blur regions are identified and deblurring is only performed in the identified regions based on the proposed weighted color interpolation. Further, we modify an Android camera capture flow to remain in the required camera shot-to-shot time even including the proposed technique. Experimental results using an Android camera phone demonstrate the effectiveness of our proposed method.

In this paper, an interactive medical image visualization system to support medical therapy has been introduced, where 3D organ model with the corresponding medical image is visualized interactively for diagnosis and surgical planning. To show effectiveness of the proposed system, 3D liver model generated from CT data has been utilized in consideration of its deformable characteristics by respiration as well as appearance. In addition, a hand gesture interface is applied on the graphical user interface for providing more natural and intuitive interactivity.

This paper suggests a novel technology that can clone a user's facial expression to his avatar in 3D virtual worlds realistically using only one color camera on a smart TV. To do this, the user's 3D head movement information and 3D position of facial feature points are needed in real time. We propose two novel approaches to achieve this. Firstly, we use a personalized 3D and 2D facial expression model to deal with head movement and various expressions. Secondly, we use a facial muscle model to generate natural motion of facial feature points located in cheeks and forehead which are difficult to be tracked using a camera. Experimental results demonstrate that the proposed method would be an efficient technique to perform realistic 3D facial expression cloning.

Respiratory motion monitoring is important to deal with tumor tracking in radiotherapy and imaging. In this paper, we propose a system that measures directly the motion of organ itself only using ultrasound image. Our system locates robustly diaphragm even though diaphragm may be blurred in ultrasound images by a fitting technique. In addition, it can extract robustly respiratory signal by automatically selecting the best area to monitor respiratory area in all image sequences.

Image-guided therapy is the treatment technology which uses medical images to understand anatomic information on human body and make medical treatment plan. Such treatment includes radio frequency (RF) ablation, high intensity focused ultrasound (HIFU), cyber knife, surgical robot, etc. Especially, ultrasound (US) guided HIFU therapy has attracted much attention since it treats internal organs without giving any radiation dose to human body. However, it is still challengeable to track 3D position of tumors of moving organ from 2D us videos since respiration makes change on the position and shape of the internal organs such as liver, kidney or pancreas. The paper proposes a moving organ tracking method based on 3D organ model. The 3D organ model is made from MR or CT images captured before treatment. The 3D organ model is registered to US images, and the organ and its tumor's position are tracked in 2D US videos in real time. We apply the method on a liver phantom and get 5.3ms as the processing time and 1.3 mm as the accuracy.

Realizing the very natural video by using holography technology is thought as the ultimate goal of 3D display, and it is still remained as the Holy Grail to be explored. Difficulties lie on two challenging issues: development of the Giga-pixel holographic video display which is feasible and meaningful for commercialization and fast calculation algorithms for the real-time holographic video generation. This presentation reviews the current status in holographic video generation methods in the aspect of the latter issue. The recent algorithms can be categorized into two main concepts of the point source hologram and the Fourier transform hologram. By analyzing the complexity of the algorithms, we realized that there is a common limitation regardless of the type of algorithms. I will address this issue and discuss the research directions to move one step forward for the commercial holographic video displays.

The framework of capturing multiview images has received a great attention from industry as well as academy due to its usability in various applications; 3D scene reconstruction, super-resolution, scene understanding, surveillance etc. This paper presents a multiview capturing system that uses a single camera with multiple mirrors. We place multiple mirrors pointing to the target scene and record an image of those mirrors. From this image of mirrors, we can extract the multiview images of the target scene, images reflected by mirrors. Based on the proposed method, we achieve some success in geometry reconstruction and view synthesis using multiview images, which shows the usability of the proposed work. The main contribution of this work is to introduce a light weight implementation for multiview capturing with a number of useful properties; the automatic synchronization across multiple views, the consistent color/white balance setting across multiple views, the capability of recording a wide range of view and the cost-effective implementation.

In this paper, we propose a novel mobile LCD device for simultaneous touch and hover sensing. A transparent infrared image sensor panel allows the device to display contents and to detect touch and hover actions alternately. The proposed system has an advantage over touch screen devices in that a user interacts with a mobile device by not only touching the display screen but also hovering over the display screen.

This paper presents a new 6 DOF motion tracking architecture which can be utilized in wide application area. Based on spatial IR intensity attenuation characteristics, 3D position and orientation of the device are estimated accurately in real-time, and its performance is more improved by inertial sensor fusion. The proposed system can be constructed with low-cost sensors, and it has advantageous architecture for downsizing without any deterioration of its performance.

Multiview video sequences with more than 3 color videos and additional supplementary data require huge amount of transmission and storage capacities. This paper proposes novel approaches to transfer such multiview video contents more efficiently by applying a mixed resolution video format and new compression schemes utilizing the characteristics of multiview video contents.

Thermoelectric material bismuth telluride (Bi 2 Te 3 ) nanowires were synthesized by pulsed-potential deposition using Bi(NO 3 ) 3 ·5H 2 O and TeO 2 in 1.0 M HNO 3 in porous anodic alumina (PAA) template. Nanowires with various surface morphologies were obtained by changing the potential between −0.2 and −0.6 V vs. SCE (sat. KCl). The more negative pulse potentials than the reduction potential observed from cyclic voltammograms (CVs) were used for nanowires growth, delay potential was used in no current region between reduction and oxidation potentials of the compound. To obtain dense, highly oriented single-crystalline nanowires, the control of the growth rate is critical. Surface roughness and crystallinity of nanowires were affected by the pulsed-potential, and the pulsed-time. The microstructures and morphologies of nanowires were investigated by powder X-ray diffraction, FE-SEM, and HR-TEM. Thermoelectric properties of both smooth and rough surfaced nanowires are under investigation.

We present a bi-layer inpainting method for synthesizing novel views from a single color image and its corresponding depth map under the exemplar-based inpainting framework. Unlike conventional image inpainting, the decisions of the inpainting direction and the sample regions are important to inpaint disoccluded regions which disclose hidden background regions in the new viewpoint. The proposed algorithm first labels boundaries along the disoccluded regions whether it belongs to the foreground or background objects, and then separates their surrounding regions into the foreground and background regions using the graph cut algorithm. The disoc-cluded regions are then filled from the background boundary with best-match patches taken from the background regions. As demonstrated in the experimental results, the proposed method recovers the disoccluded region with visually plausible quality.

Nonlocal means filtering is an edge-preserving denoising method whose filter weights are determined by Gaussian weighted patch similarities. The nonlocal means filter shows superior performance in removing additive Gaussian noise at the expense of high computational complexity. In this paper, we propose an efficient and effective denoising method by introducing a separable implementation of the nonlocal means filter and adopting a bilateral kernel for computing patch similarities. Experimental results demonstrate that the proposed method provides comparable performance to the original nonlocal means, with lower computational complexity.

A three-dimensional (3D) mesh compression algorithm based on novel prediction methods and a mode decision scheme is proposed in this work. After decomposing an input mesh into base and refinement layers, we segment the geometry data of each layer into clusters. To encode vertex positions efficiently, we propose two prediction methods: the dual ring prediction and the minimum mean square error (MMSE) prediction. Also, we develop a mode decision scheme that selects the best prediction mode for each cluster. Simulation results demonstrate that the proposed algorithm provides significantly better compression performance than conventional techniques.

Text plagiarism is growing rapidly with the development of Internet, so many plagiarism detection algorithms have been proposed. However, most algorithms focus on the optimized one-to-one comparison, rather than massive document comparison. The latter algorithms have a limitation in time performance when users conduct an exhaustive search on a huge set of documents. In this paper, we propose an optimized preprocessing model to detect similar text in massive document repositories. This model uses an efficient data structure called GDIC (Global Dictionary) for preprocessing. After filtering stop words, we choose pairs of documents to be inspected using two methods at the same time, both of which use the concept of a common non-stop word to choose pairs of documents to be inspected, each of which uses it in a slightly different way. The first method chooses pairs of documents with a high frequency of common non-stop words in documents in each of these pairs, while the second method chooses pairs with a high proportion of common non-stop words. We experimentally prove the performance of the model. Our experiments with the proposed preprocessing model is drastically reduced searching time to 64~87%, while the sensitivity stands at 77~96%. When we use this model, GDIC generation time accounts for a large proportion of all of the detection time. In future work, we will optimize GDIC creation time to improve the performance of the entire system.

A novice user is generally required to pass the Human Interactive Proof(HIP) before they are allowed to register web services, since there are so many automated software tools and automated scripts available in public web services. Thus there are many variations to help HIP. Text-based CAPTCHA is the most common method, where each user is asked to discern the hidden characters in noisy-images. Many image-based HIP were recently proposed and tested to overcome the weakness of text-based HIP. Though the image-based HIP problem is harder to solve compared to a conventional text-based puzzle, some powerful machine learning tools, such as Support Vector Machine (SVM), can easily break this problem. The authors have previously developed an image orientation-based CAPTCHA problem with polygonally cropped sub-images. In this paper, we address the more fundamental issue: what is the most important feature for a human to recognize the correct orientation of a partial image. We insist that the box-counting measure for an arbitrary polygon, which was already been devised to measure fractal objects, is a successful metric to measure the human perception of partial image orientation. This means that it is - easy for human to get the orientation of the high-box counted polygonal sub-image, but very hard for any machine learning tool due to the partialness of subject images. We also insist that a special polygon which has a high box-counting value, is a fractal shape, so an effective partial image for HIP is an empirically random fractal polygon. We found that irregularly recurring fractal polygons defeat SVM attack to decrease random attack accuracy.

The 3DTV market has already started and more people have been enjoying new visual experience. But annoyance of wearing 3D glasses is a problem and the autostereoscopic display still has several limitations unsolved. This presentation discusses these limitations and our approaches to overcome them.

The multi-view video coding (MVC) standard includes all the highly advanced techniques of H.264/AVC. Especially, the techniques employed in motion estimation are extended to reduce not only temporal but inter-view redundancies in MVC, but motion vector prediction of a skip mode is only performed in a temporal direction. Therefore, we introduce a temporal skip mode using reference pictures at the same view as the current picture and an inter-view skip mode employing reference pictures at the different views from the current picture. The proposed method decides the best one of the temporal and the inter-view skip modes, based on rate-distortion (RD) optimization. Experimental results illustrate that the proposed method is significantly more efficient than the conventional method.

The purpose of this project is to take a lead in the environmental protection and build an environmentally friendly image by using eco-bags as a development tool for medical tour packages. Two designs are discussed here. The strengths of an eco-bag include being lightweight, portable, and environmentally friendly in place of a plastic bag. With the design, the cold image of existing hospitals may be transformed anew.

The recent online networked virtual worlds such as SecondLife, World of Warcraft and Lineage have been increasingly popular. A life-scale virtual world presentation and the intuitive interaction between the users and the virtual worlds would provide more natural and immersive experience for users. The emergence of novel interaction technologies such as sensing the facial expression and the motion of the users and the real world environments could be used to provide a strong connection between them. For the wide acceptance and use of the virtual world, a various type of novel interaction devices should have a unified interaction formats between the real world and the virtual world and interoperability among virtual worlds. Thus, MPEG-V Media Context and Control (ISO/IEC 23005) standardizes such connecting information. The paper provides an overview and its usage example of MPEG-V from the real world to the virtual world (R2V) on interfaces for controlling avatars and virtual objects in the virtual world by the real world devices. In particular, we investigate how the MPEG-V framework can be applied for the facial animation of an avatar in various types of virtual worlds.

In this work, we propose a progressive compression algorithm using topology-based Karhunen-Loeve transform(KLT). First, we simplify an input mesh to represents an original mesh in several level of details. Then, coordinates of decimated vertices at each level are predicted from the coarser level mesh, and the prediction residuals are transmitted to a decoder. To provide high coding efficiency, we apply the topology-based KLT, which compacts the energy into a few coefficients, to the prediction residuals. Moreover, we develop a bit plane coder, which uses a context-adaptive arithmetic coder, for the entropy coding. Experiments on various 3D meshes show that the proposed algorithm provides enhanced compression performance.

In this paper, we present a new video coding scheme to improve coding efficiency. For this purpose, we develop a loop filter that restores the compressed pictures in accordance with the image local characteristic. More specifically, a picture is segmented in such a way that the rate distortion cost is minimized and each pixel of the picture is restored using Wiener filter specific to the region. The restored picture is used for the motion-compensated prediction of future pictures and this can improve coding efficiency. In particular, we propose segmentation mode to reflect the local attributes of a picture. Finally, video coding that performs the new approach is implemented and results in 12.7% BD-bitrate reduction on average in Baseline Profile when compared to H.264/AVC.

This paper presents a novel algorithm for realistic 3D face modeling from a single pair of color and time-of-flight (TOF) depth images using a generic, deformable model. Most previous approaches have attempted to emphasize either facial feature or global shape consistency, resulting in either visually plausible or geometrically accurate models. In this paper, we introduce feature-preserving surface registration to achieve both reality and accuracy. The proposed algorithm takes advantages of both facial features from the color image and geometric information from the depth image in an iterative closest point (ICP) framework, where the non-rigid registration is achieved by iteratively minimizing the energy function consisting of feature distance term, surface distance term, and linear elasticity term. As demonstrated in the experimental results, the proposed algorithm builds the personalized 3D face model within a few seconds with high visual quality.

Time-of-Flight depth cameras provide a direct way to acquire range images, using the phase delay of the incoming reflected signal with respect to the emitted signal. These cameras, however, have a challenging problem called range folding, which occurs due to the modular error in phase delay—ranges are modulo the maximum range. To our best knowledge, we exploit the first approach to estimate the number of mods at each pixel from only a single range image. The estimation is recasted into an optimization problem in the Markov random field framework, where the number of mods is considered as a label. The actual range is then recovered using the optimal number of mods at each pixel, so-named range unfolding. As demonstrated in the experiments with various range images of real scenes, the proposed method accurately determines the number of mods. In result, the maximum range is practically extended at least twice of that specified by the modulation frequency.

This paper presents a novel approach to face modeling for realistic synthesis, powered by a probabilistic face diffuse model and a generic face specular map. We first construct a probabilistic face diffuse model for estimating the albedo and the normals of a face from an unknown input image. Then, we introduce a generic face specular map for estimating the specularity of the face. Using the estimated albedo, normal and specular information, we can synthesize the face under arbitrary lighting and viewing directions realistically. Unlike many existing face modeling techniques, our approach can retain both the diffuse and specular properties of the face without involving an elaborating 3D matching procedure. Thanks to the compact representation and the effective inference scheme, our technique can be applied to many practical applications, such as face normalization, avatar creation and de-identification.