Search

This paper tackles a challenging problem of inertial sensor-based recognition for similar gait action classes (such as walking on flat ground, up/down stairs, and up/down a slope). We solve three drawbacks of existing methods in the case of gait actions: the action signal segmentation, the sensor orientation inconsistency, and the recognition of similar action classes. First, to robustly segment the walking action under drastic changes in various factors such as speed, intensity, style, and sensor orientation of different participants, we rely on the likelihood of heel strike computed employing a scale-space technique. Second, to solve the problem of 3D sensor orientation inconsistency when matching the signals captured at different sensor orientations, we correct the sensor?s tilt before applying an orientation-compensative matching algorithm to solve the remaining angle. Third, to accurately classify similar actions, we incorporate the interclass relationship in the feature vector for recognition. In experiments, the proposed algorithms were positively validated with 460 participants (the largest number in the research field), and five similar gait action classes (namely walking on flat ground, up/down stairs, and up/down a slope) captured by three inertial sensors at different positions (center, left, and right) and orientations on the participant?s waist. HighlightsAn action recognition algorithm for similar gait actions using an inertial sensor.A robust period segmentation based on the likelihood of an heel-strike is presented.The proposed method works well against a variation of sensor orientation.Interclass relationship improves the recognition accuracy significantly.The accuracy is more than 91% for a very large database of 460 subjects.

The goal of image annotation is to automatically assign a set of textual labels to an image to describe the visual contents thereof. Recently, with the rapid increase in the number of web images, nearest neighbor (NN) based methods have become more attractive and have shown exciting results for image annotation. One of the key challenges of these methods is to define an appropriate similarity measure between images for neighbor selection. Several distance metric learning (DML) algorithms derived from traditional image classification problems have been applied to annotation tasks. However, a fundamental limitation of applying DML to image annotation is that it learns a single global distance metric over the entire image collection and measures the distance between image pairs in the image-level. For multi-label annotation problems, it may be more reasonable to measure similarity of image pairs in the label-level. In this paper, we develop a novel label prediction scheme utilizing multiple label-specific local metrics for label-level similarity measure, and propose two different local metric learning methods in a multi-task learning (MTL) framework. Extensive experimental results on two challenging annotation datasets demonstrate that 1) utilizing multiple local distance metrics to learn label-level distances is superior to using a single global metric in label prediction, and 2) the proposed methods using the MTL framework to learn multiple local metrics simultaneously can model the commonalities of labels, thereby facilitating label prediction results to achieve state-of-the-art annotation performance.

A thermal camera captures the temperature distribution of a scene as a thermal image. In thermal images, facial appearances of different people under different lighting conditions are similar. This is because facial temperature distribution is generally constant and not affected by lighting condition. This similarity in face appearances is advantageous for face detection. To detect faces in thermal images, cascade classifiers with Haar-like features are generally used. However, there are few studies exploring the local features for face detection in thermal images. In this paper, we introduce two approaches relying on local features for face detection in thermal images. First, we create new feature types by extending Multi-Block LBP. We consider a margin around the reference and the generally constant distribution of facial temperature. In this way, we make the features more robust to image noise and more effective for face detection in thermal images. Second, we propose an AdaBoost-based training method to get cascade classifiers with multiple types of local features. These feature types have different advantages. In this way we enhance the description power of local features. We did a hold-out validation experiment and a field experiment. In the hold-out validation experiment, we captured a dataset from 20 participants, comprising 14 males and 6 females. For each participant, we captured 420 images with 10 variations in camera distance, 21 poses, and 2 appearances (participant with/without glasses). We compared the performance of cascade classifiers trained by different sets of the features. The experiment results showed that the proposed approaches effectively improve the performance of face detection in thermal images. In the field experiment, we compared the face detection performance in realistic scenes using thermal and RGB images, and gave discussion based on the results.

A light field, which is often understood as a set of dense multi-view images, has been utilized in various 2D/3D applications. Efficient light field acquisition using a coded aperture camera is the target problem considered in this paper. Specifically, the entire light field, which consists of many images, should be reconstructed from only a few images that are captured through different aperture patterns. In previous work, this problem has often been discussed from the context of compressed sensing (CS). In contrast, we formulated this problem from the perspective of principal component analysis (PCA) to derive optimal non-negative aperture patterns and a straight-forward reconstruction algorithm. Even though it is based on a conventional technique, our method has proven to be more accurate and much faster than a state-of-the-art CS-based method.

This paper presents an incremental structural modeling approach that improves the precision and the stability of existing batch-based ones for sparse and noisy point clouds from visual simultaneous localization and mapping (SLAM). The main idea is to use the generating process of point clouds on SLAM effectively. First, a batch-based method is applied to point clouds that are incrementally generated from SLAM. Then, the temporal history of reconstructed geometric primitives is statistically analyzed to suppress incorrect reconstruction. The evaluation shows that both precision and stability are improved compared to an existing batch-based method, and the proposed method is suitable for real-time structural modeling.

An infrared (IR) camera captures the temperature distribution of an object as an IR image. Because facial temperature is almost constant, an IR camera has the potential to be used in detecting facial regions in IR images. However, in detecting faces, a simple temperature thresholding does not always work reliably. The standard face detection algorithm used is AdaBoost with local features, such as Haar-like, MB-LBP, and HoG features in the visible images. However, there are few studies using these local features in IR image analysis. In this paper, we propose an AdaBoost-based training method to mix these local features for face detection in thermal images. In an experiment, we captured a dataset from 20 participants, comprising 14 males and 6 females, with 10 variations in camera distance, 21 poses, and participants with and without glasses. Using leave-one-out cross-validation, we show that the proposed mixed features have an advantage over all the regular local features.

Hyperspectral imaging is used in various fields because it can obtain much more information than imaging by conventional RGB cameras. Hyperspectral imaging systems using active illumination, prisms, gratings, or narrowband filters have been proposed. Active illumination systems can obtain two-dimensional (2D) spectral images rapidly, and the device can be low-cost and small because of the use of LEDs. However, flicker can occur when different colors of LEDs are switched. The other methods do not have the flicker problem because they use passive imaging. However, these systems take a long time to acquire the 2D spectral images, or they tend to be high-cost or large. In our research, we propose a flickerless active LED illumination system for hyperspectral imaging. This system acquires images while switching the illumination. The switching illumination consists of many narrowband LEDs that have different spectrums. The spectral images of each LED are reconstructed from the acquired images. The switching illumination is designed to reduce the flicker based on human visual characteristics. We reduce the color changes of the switching illumination while maintaining its spectral differences. In the experiment, we obtain the optimal design of a flickerless illumination system for measuring oxygen saturation. To show the feasibility of our system, we clearly show the difference in saturation using the spectral images obtained by a prototype designed using the proposed method.

Several recent studies in compressive video sensing have realized scene capture beyond the fundamental trade-off limit between spatial resolution and temporal resolution using random space-time sampling. However, most of these studies showed results for higher frame rate video that were produced by simulation experiments or using an optically simulated random sampling camera, because there are currently no commercially available image sensors with random exposure or sampling capabilities. We fabricated a prototype complementary metal oxide semiconductor (CMOS) image sensor with quasi pixel-wise exposure timing that can realize nonuniform space-time sampling. The prototype sensor can reset exposures independently by columns and fix these amount of exposure by rows for each 8×8 pixel block. This CMOS sensor is not fully controllable via the pixels, and has line-dependent controls, but it offers flexibility when compared with regular CMOS or charge-coupled device sensors with global or rolling shutters. We propose a method to realize pseudo-random sampling for high-speed video acquisition that uses the flexibility of the CMOS sensor. We reconstruct the high-speed video sequence from the images produced by pseudo-random sampling using an over-complete dictionary. The proposed method also removes the rolling shutter effect from the reconstructed video.

Depth from defocus (DFD) is a technique that restores scene depth based on the amount of defocus blur in the images. DFD usually captures two differently focused images, one near-focused and the other far-focused, and calculates the size of the defocus blur in these images. However, DFD using a regular circular aperture is not sensitive to depth, since the point spread function (PSF) is symmetric and only the radius changes with the depth. In recent years, the coded aperture technique, which uses a special pattern for the aperture to engineer the PSF, has been used to improve the accuracy of DFD estimation. The technique is often used to restore an all-in-focus image and estimate depth in DFD applications. Use of a coded aperture has a disadvantage in terms of image deblurring, since deblurring requires a higher signal-to-noise ratio (SNR) of the captured images. The aperture attenuates incoming light in controlling the PSF and, as a result, decreases the input image SNR. In this paper, we propose a new computational imaging approach for DFD estimation using focus changes during image integration to engineer the PSF. We capture input images with a higher SNR since we can control the PSF with a wide aperture setting unlike with a coded aperture. We confirm the effectiveness of the method through experimental comparisons with conventional DFD and the coded aperture approach. We propose a novel imaging approach, "Half-sweep imaging", for depth from defocus.Novel PSF engineering by focus changes during each exposure time.Complementary responses of the PSFs help image restoration and depth estimation.Increasing the quantity of the incident light affects the quality of DFD estimation.We built the prototype camera and confirmed the effectiveness of our approach.

We propose a method for phase estimation of a single non-parametric quasi-periodic signal. Assuming signal intensities should be equal among samples of the same phase, such corresponding samples are obtained by self-dynamic time warping between a quasi-periodic signal and a signal with multiple-period shifts applied. A phase sequence is then estimated in a sub-sampling order using an optimization framework incorporating 1) a data term derived from the correspondences and 2) a smoothness term of the local phase evolution under 3) a monotonic-increasing constraint on the phase. Such a phase estimation is, however, ill-posed because of combination ambiguity between the phase evolution and the normalized periodic signal, and hence can result in a biased solution. Therefore, we introduce into the optimization framework 4) a bias correction term, which imposes zero-bias from the linear phase evolution. Analysis of the quasi-periodic signals from both simulated and real data indicate the effectiveness and also potential applications of the proposed method.

Gait-based owner authentication using accelerometers has recently been extensively studied owing to the development of wearable electronic devices. An actual gait signal is always subject to change due to many factors including variation of sensor attachment. In this research, we tackle to the practical sensor-orientation inconsistency, for which signal sequences are captured at different sensor orientations. We present an iterative signal matching algorithm based on phase-registration technique to simultaneously estimate relative sensor-orientation and register the 3D acceleration signals. The iterative framework is initialized by using 1D orientation-invariant resultant signals which are computed from 3D signals. As a result, the matching algorithm is robust to any initial sensor-orientation. This matching algorithm is used to match a probe and a gallery signals in the proposed owner authentication method. Experiments using actual gait signals under various conditions such as different days, sensors, weights being carried, and sensor orientations show that our authentication method achieves positive results.

This paper presents the largest inertial sensor-based gait database in the world, which is made open to the research community, and its application to a statistically reliable performance evaluation for gait-based personal authentication. We construct several datasets for both accelerometer and gyroscope of three inertial measurement units and a smartphone around the waist of a subject, which include at most 744 subjects (389 males and 355 females) with ages ranging from 2 to 78 years. The database has several advantages: a large number of subjects with a balanced gender ratio, variations of sensor types, sensor locations, and ground slope conditions. Therefore, we can reliably analyze the dependence of gait authentication performance on a number of factors such as gender, age group, sensor type, ground condition, and sensor location. The results with the latest existing authentication methods provide several insights for these factors. HighlightsWe present the world largest inertial sensor-based database to the community.Based on the database, females have a better recognition performance than males.People have the best recognition performance at their twenties.An accelerometer has a better recognition performance than a gyroscope.

Background modeling and subtraction is an essential task in video surveillance applications. Many researchers have discussed about an improvement of performance of a background model, and a reduction of memory usage or computational cost. To adapt to background changes, a background model has been enhanced by introducing various information including a spatial consistency, a temporal tendency, etc. with a large memory allocation. Meanwhile, an approach to reduce a memory cost cannot provide better accuracy of a background subtraction. To tackle the trade-off problem, this paper proposes a novel framework named "case-based background modeling". The characteristics of the proposed method are (1) a background model is created, or removed when necessary, (2) case-by-case model sharing by some of the pixels, (3) pixel features are divided into two groups, one for model selection and the other for modeling. These approaches realize a low-cost and high accurate background model. The memory usage and the computational cost could be reduced by half of a traditional method and the accuracy was superior to the method.

In this paper, we propose an efficient algorithm to recognize actions of human interaction. Unlike previous algorithms using two participants' actions, the proposed algorithm estimates the action contribution of participants to determine which participant's action is the major action for correct interaction recognition. To estimate this contribution, we construct a contribution interaction model for each interaction category in which both participants carry out major actions. Using the contribution models, we design a method that automatically estimates the contribution of participants and classifies interaction samples into “co-contribution” and “single-contribution” interactions. At the same time, the major actions in the “single-contribution” interactions are determined. We evaluate our method on the UT-interaction dataset and our original interaction dataset (LIMU). Recognition results indicate the robustness of the proposed method and the high estimation accuracy obtained: estimation accuracies of 96 and 98% in set 1 and set 2 of the UT dataset, respectively, and 97.8% in the LIMU dataset. Based on the estimation results, we extract the major action information for interaction recognition. Average recognition accuracies of 93.3% in set 1 and 91.7% in set 2 of the UT dataset were obtained. Our result is at least 5% better than those obtained with previous algorithms. For the LIMU dataset, recognition accuracy reached 91.1%. It was 8.9% higher than the recognition result without contribution estimation. © 2013 Institute of Electrical Engineers of Japan. Published by John Wiley & Sons, Inc.

In these days, "early recognition" of gesture patterns has been studied by many researchers. Early recognition is a method to make a decision of gesture recognition at the beginning part of it. In traditional method, the key postures for a gesture are utilized for recognition and early recognition is performed frame-by-frame. However, this method has a problem that computational time in recognition processing increases in proportion to size of posture database. If the processing time becomes longer, some input frames will be ignored from the processing. It results in lower recognition accuracy. In this paper, we introduce a hash-based approach to search the posture database. It realizes real-time processing, and keep high performance of recognition.

We propose a method for forecasting power demand based on action plan declaration. Conventional method uses sensing data such as time series of power consumption and weather information and fails to forecast power demand due to unstable user behavior. By using action plan declared by user, meanwhile, we achieve rough forecasting of power demand.

In this paper, we propose a new method to calculate local features. We extend the FAST corner detector to the spatiotemporal space to extract the shape and motion information of human actions. And a compact peak-kept histogram of oriented spatiotemporal gradients (CHOG3D) is proposed to calculate local features. CHOG3D is calculated in a small support region of a feature point, and it employs the simplest gradient, the first-order gradient, for descriptor calculation. Through parameter training, the proper length of the CHOG3D is determined to be 80 elements, which is 1/12.5 times the dimension of HOG3D in the KTH dataset. In addition, it keeps the peak value of quantized gradient to represent human actions more exactly and distinguish them more correctly. CHOG3D overcomes the disadvantages of the complex calculation and huge length of the descriptor HOG3D. From a comparison of the computation cost, CHOG3D is 7.56 times faster than HOG3D in the KTH dataset. We apply the algorithm for human action recognition with support vector machine. The results show that our method outperforms HOG3D and some other currently used algorithms. © 2012 Institute of Electrical Engineers of Japan. Published by John Wiley & Sons, Inc.

Since 1960s, aperture patterns have been studied extensively and a variety of coded apertures have been proposed for various applications, including extended depth of field, defocus deblurring, depth from defocus, light field acquisition, etc. Researches have shown that optimal aperture patterns can be quite different due to different applications, imaging conditions, or scene contents. In addition, many coded aperture techniques require aperture patterns to be temporally changed during capturing. As a result, it is often necessary to have a programmable aperture camera whose aperture pattern can be dynamically changed as needed in order to capture more useful information. In this paper, we propose a programmable aperture camera using a Liquid Crystal on Silicon (LCoS) device. This design affords a high brightness contrast and high resolution aperture with a relatively low light loss, and enables one change the pattern at a reasonably high frame rate. We build a prototype camera and evaluate its features and drawbacks comprehensively by experiments. We also demonstrate three coded aperture applications in defocus deblurring, depth from defocus and light field acquisition.

Object detection is an important task for computer vision applications. Many researchers have proposed a number of methods to detect the objects through background modeling. To adapt to “illumination changes” in the background, local feature-based background models are proposed. They assume that local features are not affected by background changes. However, “motion changes”, such as the movement of trees, affect the local features in the background significantly. Therefore, it is difficult for local feature-based models to handle motion changes in the background. To solve this problem, we propose a new background model in this paper by applying a statistical framework to a local feature-based approach. Our proposed method combines the concepts of statistical and local feature-based approaches into a single framework. In particular, we use illumination invariant local features and describe their distribution by Gaussian Mixture Models (GMMs). The local feature has the ability to tolerate the effects of “illumination changes”, and the GMM can learn the variety of “motion changes”. As a result, this method can handle both background changes. Some experimental results show that the proposed method can detect the foreground objects robustly against both illumination changes and motion changes in the background.

For fast egomotion of a camera, computing feature correspondence and motion parameters by global search becomes highly time-consuming. Therefore, the complexity of the estimation needs to be reduced for real-time applications. In this paper, we propose a compound omnidirectional vision sensor and an algorithm for estimating its fast egomotion. The proposed sensor has both multi-baselines and a large field of view (FOV). Our method uses the multi-baseline stereo vision capability to classify feature points as near or far features. After the classification, we can estimate the camera rotation and translation separately by using random sample consensus (RANSAC) to reduce the computational complexity. The large FOV also improves the robustness since the translation and rotation are clearly distinguished. To date, there has been no work on combining multi-baseline stereo with large FOV characteristics for estimation, even though these characteristics are individually are important in improving egomotion estimation. Experiments showed that the proposed method is robust and produces reasonable accuracy in real time for fast motion of the sensor.

The latest robust estimators usually take advantage of density estimation, such as kernel density estimation, to improve the robustness of inlier detection. However, the challenging problem for these systems is choosing the suitable smoothing parameter, which can result in the population of inliers being overor under-estimated, and this, in turn, reduces the robustness of the estimation. To solve this problem, we propose a robust estimator that estimates an accurate inlier scale. The proposed method first carries out an analysis to figure out the residual distribution model using the obvious case-dependent constraint, the residual function. Then the proposed inlier scale estimator performs a global search for the scale producing the residual distribution that best fits the residual distribution model. Knowledge about the residual distribution model provides a major advantage that allows us to estimate the inlier scale correctly, thereby improving the estimation robustness. Experiments with various simulations and real data are carried out to validate our algorithm, which shows certain benefits compared with several of the latest robust estimators.

The segmentation of transparent objects can be very useful in computer vision applications. However, because they borrow texture from their background and have a similar appearance to their surroundings, transparent objects are not handled well by regular image segmentation methods. We propose a method that overcomes these problems using the consistency and distortion properties of a light-field image. Graph-cut optimization is applied for the pixel labeling problem. The light-field linearity is used to estimate the likelihood of a pixel belonging to the transparent object or Lambertian background, and the occlusion detector is used to find the occlusion boundary. We acquire a light field dataset for the transparent object, and use this dataset to evaluate our method. The results demonstrate that the proposed method successfully segments transparent objects from the background., 9 pages, 14 figures, 2 tables, ICCV 2015

International audience; Image-based modeling methods for generating 3D models from an image sequence have been widely studied. Most of these methods, however, require huge redundant spatio-temporal images to estimate scene depth. This is not an effective use of capturing higher resolution texture. On the other hand, a route panorama, which is a continuous panoramic image along a path, is an efficient way of consolidating information from multiple viewpoints into a single image. A route panorama captured by a line camera also has the advantage of capturing higher resolution easily. In this paper, we propose a method for estimating the depth of an image from a route panorama using color drifts. The proposed method detects color drift by deformable window matching of the color channels. It also uses a hierarchical belief propagation to estimate the depth stably and decrease the computation cost thereof.

In this tutorial, we will introduce emerging technologies on computational imaging and light field projection to AR/MR researchers.Light is the most important medium in AR/VR technologies to not only obtain information but also show and modify visual cue in the real scenes. Therefore in this area, latest techniques on optics, imaging and lighting have played an important role to make a next step toward the sophisticated experiences. Computational photography is one of the most influential technology in computer vision and optical engineering areas, and we think most techniques in computational imaging and projection can be applied to common problems in mixed reality, such as scene modeling, modification of the appearances of actual objects and user interactions.

In this paper, we present a visualization tool for person re-identification when tracking objects across non-overlapping cameras. Tracking objects across non-overlapping cameras is challenging because the observations from different cameras are widely separated in both time and space. Hence, these systems need a large amount of labeled training data. Commonly, this training data is constructed manually at significant human cost. We support this process efficiently by visualizing the correspondences of objects across multiple cameras. Our tool facilitates the construction of a database for person re-identification with ease. Moreover, the accuracy of person re-identification can be increased using the generated database because the amount of training data is increased. In the experiments, we apply the proposed tool to real world situations to verify the validity of the proposed system.

We introduce a method to recover the shape of a smooth dielectric object from polarization images taken with a light source from different directions. We present two constraints on shading and polarization and use both in a single optimization scheme. This integration is motivated by the fact that photometric stereo and polarization-based methods have complementary abilities. The polarization-based method can give strong cues for the surface orientation and refractive index, which are independent of the light direction. However, it has ambiguities in selecting between two ambiguous choices of the surface orientation, in the relationship between refractive index and zenith angle (observing angle), and limited performance for surface points with small zenith angles, where the polarization effect is weak. In contrast, photometric stereo method with multiple light sources can disambiguate the surface orientation and give a strong relationship between the surface normals and light directions. However, it has limited performance for large zenith angles, refractive index estimation, and faces the ambiguity in case the light direction is unknown. Taking their advantages, our proposed method can recover the surface normals for both small and large zenith angles, the light directions, and the refractive indexes of the object. The proposed method is successfully evaluated by simulation and real-world experiments.

Making a reasonable ranking on images in dataset is one of the main objectives for object retrieval challenge, and in this paper we intend to improve the ranking quality. We follow the idea of query expansion in previous researches. Based on the use of bag-of-visual-words model, tf-idf scoring and spatial verification, previous method applied a pointwise style learning in query expansion stage, using but not fully exploring verification results. We intend to extend their learning approach for better discriminative power in retrieval. In re-ranking stage we propose a method using pairwise learning, instead of pointwise learning previously used. We could obtain more reliable ranking on a shortlist of examples. If this verification itself is reliable, a good re-ranking should best preserve this sub-ranking order. Thus in our proposed method, we are motivated to leverage a pairwise learning method to incorporate the ranking sequential information more efficiently. We evaluate and compare our proposed method with previous methods over Oxford 5k dataset, a standard benchmark dataset, where our method achieve better mean average precision and showed better discriminative power.

Since 1960s, aperture patterns have been studied extensively and a variety of coded apertures have been proposed for various applications, including extended depth of field, defocus deblurring, depth from defocus, light field acquisition, etc. Researches have shown that optimal aperture patterns can be quite different due to different applications, imaging conditions, or scene contents. In addition, many coded aperture techniques require aperture patterns to be temporally changed during capturing. As a result, it is often necessary to have a programmable aperture camera whose aperture pattern can be dynamically changed as needed in order to capture more useful information. In this paper, we propose a programmable aperture camera using a Liquid Crystal on Silicon (LCoS) device. This design affords a high brightness contrast and high resolution aperture with a relatively low light loss, and enables one change the pattern at a reasonably high frame rate. We build a prototype camera and evaluate its features and drawbacks comprehensively by experiments. We also demonstrate four coded aperture applications in defocus deblurring, depth from defocus, light field acquisition, and motion deblurring.

Detection of moving objects is one of the key steps for vision based applications. Many previous works leverage background subtraction using background models and assume that image sequences are captured from a stationary camera. These methods are not directly applied to image sequences from a moving camera because both foreground and background objects move with respect to the camera. One of the approaches to tackle this problem is to estimate background movement by computing pixel correspondences between frames such as homography. With this approach, moving objects can be detected by using existing background subtraction. In this paper, we evaluate detection of foreground objects for image sequences from a moving camera. Especially, we focus on homography as a camera motion. In our evaluation we change the following parameters: changing feature points, the number of them and estimation methods of homography. We analyze its effect on detection of moving objects in regard to detection accuracy, processing time. Through experiments, we show requirement of background models in image sequences form a moving camera.

Many virtual reality, mixed reality, and telepresence applications use head mounted displays (HMD). HMD systems are portable and can display stereoscopic images. However, the field of view (FOV) of commercial HMD systems is too narrow for conveying the feeling of immersion. The horizontal FOV is typically around 60°, significantly narrower than that of the human eye. In this paper, we propose new display optics for a super wide FOV head mounted display. The proposed optics consists of an ellipsoidal and a hyperboloidal mirror that will display distortionless images by using the characteristics of the mirrors, even if the image has a large FOV. We constructed a prototype HMD system with a 180° horizontal × 60° vertical FOV that includes the peripheral vision of the human eye. The FOV has a 60° × 60° overlap area that can display stereoscopic images. We estimated the resolution, focus, and aberration of the prototype in an optical simulation and experimentally confirmed that the prototype displays distortionless wide FOV images.

We propose a novel strategy to obtain a high spatio-temporal resolution video. To this end, we introduce a dual sensor camera that can capture two video sequences with the same field of view simultaneously. These sequences record high resolution with low frame rate and low resolution with high frame rate. This paper presents an algorithm to synthesize a high spatio-temporal resolution video from these two video sequences by using motion compensation and spectral fusion. We confirm that the proposed method improves the resolution and frame rate of the synthesized video.

The head-mounted display (HMD) has been used widely as a visual presentation device in virtual reality, composite reality, and telepresence. However, the HMDs which are on the market have only a very narrow angular field of view, which is less than 60° in the horizontal direction. This is very narrow compared to the human field of view and is far from generating a feeling of presence. The poor feeling of presence is considered to be a factor that degrades the sense of realism for the user and the operating efficiency in the virtual space. In this paper, we propose an HMD with a wide field of view based on a catadioptrical system composed of hyperboloidal and ellipsoidal mirrors. The proposed HMD achieves a horizontal view angle of 180°, which covers the human peripheral field of vision. This wide field of view gives the user a stronger feeling of reality and ambience. The proposed HMD has a binocular overlap area of 60° from the center, which helps to achieve stereoscopic vision by presenting parallax images. © 2006 Wiley Periodicals, Inc. Electron Comm Jpn Pt 2, 89(9): 33–43, 2006; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/ecjb.20304

Recently, many virtual reality and robotics applications have been called on to create virtual environments from real scenes. A catadioptric omnidirectional image sensor composed of a convex mirror can simultaneously observe a 360-degree field of view making it useful for modeling man-made environments such as rooms, corridors, and buildings, because any landmarks around the sensor can be taken in and tracked in its large field of view. However, the angular resolution of the omnidirectional image is low because of the large field of view captured. Hence, the resolution of surface texture patterns on the three-dimensional (3-D) scene model generated is not sufficient for monitoring details. To overcome this, we propose a high resolution scene texture generation method that combines an omnidirectional image sequence using image mosaic and superresolution techniques.

In this paper we propose a new light transport refocusing method for depth estimation as well as for investigation inside scattering media with unknown scattering properties. Propagated visible light rays through scattering media are utilized in our proposed refocusing method. We use 2D light source to illuminate the scattering media and 2D image sensor for capturing transported rays. The proposed method that uses 4D light transport can clearly visualize shallow depth, as well as deep depth plane of the medium. We apply our light transport refocusing method for depth estimation using conventional depth-from-focus method and for clear visualization by descattering the light rays passing through the medium. To evaluate the effectiveness we have done experiments using acrylic and milk-water type scattering medium in various optical and geometrical conditions. Finally, we show up the results of depth estimation and clear visualization, as well as with numeric evaluation., 2014 22nd International Conference on Pattern Recognition,Stockholm, Sweden,24-28 Aug. 2014

The omnidirectional image sensor named HyperOmniVision, is composed of a hyperboloidal mirror and conventional video camera. It can observe a 360 degree field of view and can transform an input image to a perspective image. However, it has an intrinsic problem where the image resolution of HyperOmniVision is lower than that of an ordinary video camera, because it has a structure whereby only one CCD captures whole periphery scene. We propose a resolution improvement method using sub-pixel displaced and multi-focused images.

Novel direct reactions between trioxane and ethylene oxide were discovered, and three novel cyclic formals were isolated and identified. These novel cyclic compounds clarified the initiation mechanism of the copolymerization of trioxane and ethylene oxide. This type of direct reaction is not only limited to the reaction between trioxane and ethylene oxide, but it was also found to be generalized to the reaction between the cyclic formal and ethylene oxide.

The copolymerization of trioxane and ethylene oxide in the bulk polymerization state using highly purified trioxane as the starting material, boron trifluoride dibutyl ether as the initiator, and methylal as the chain-transfer agent was investigated for the industrialization of the acetal copolymer, and the various important polymerization conditions were checked. The acetal exchange reaction during the copolymerization gave a uniform acetal copolymer with randomized comonomer distribution. To obtain the polymer with a high degree of stability, several checks were made. The important points were the high purification of the monomer and complete deactivation of the catalyst residue. Thus, the polymer as polymerized showed a degree of stability over 99.5%.

In a previous report, we noted a new direct reaction between 1 mole of trioxane and 1 mole of ethylene oxide and the formation of a novel cyclic formal of 1, 3, 5, 7-tetraoxacyclononane. We have now determined a new reaction and novel cyclic compound from the reaction of trioxane and ethylene oxide. The novel cyclic compound was separated and collected using a micro-distillation apparatus and gas chromatographic method. Its chemical structure was confirmed using 1H-NMR, 13C-NMR, mass spectrum and elemental analysis. The new isolated novel cyclic compound was 1, 3, 5, 7, 10-pentaoxacyclododecane, which might be formed by the direct reaction of 1 mole of trioxane and 2 moles of ethylene oxide. In the commercial base acetal copolymer, which was produced by the copolymerization of trioxane and ethylene oxide, it is known that there is a consecutive two units sequence of ethylene oxide, and this consecutive 2 ethylene oxide unit sequence might be postulated to come from the ring opening-polymerization of the novel cyclic formal of 1, 3, 5, 7, 10-pentaoxacyclododecane.

ベンゼンの部分水素化反応のシクロヘキセン選択性を向上させるために, アルコール類の助触媒効果について検討を行った. 反応は水, もしくは金属塩水溶液中に触媒を分散させて行い, 触媒として, Ru/La (OH) 3, Ru/γ-Al2O3, Ru金属触媒を用いた. いずれの触媒を使用した場合においても,わずかな量のアルコール類の添加により反応速度の低下とともに選択性が向上し, アルコール類が助触媒として有効に働くことがわかった.アルコールの種類によって添加量と選択性向上効果の関係は大きく異なり, ブチルアルコール類を用いた場合は, 第一級アルコールが最も効果的で, 第二級第三級では効果は小さくなった. またアリルアルコール類やベンジルアルコールはごく微量の添加で反応速度を著しく阻害した. またアルコールの添加効果の程度は触媒によって異なり, Ru金属触媒を使用した系では, ベンジルアルコールを添加することにより, ベンゼン転化率30%において, シクロヘキセン選択率70%以上が得られた.これらの効果は, アルコール酸素の不対電子から触媒であるRuへの電子供与が起こり, アルコール類が配位してRu表面の電子状態を変化させるか, もしくは単に触媒表面を被覆することによって生じたものと考えられる. その結果, ベンゼンや生成シクロヘキセンの触媒への吸着や, シクロヘキサンへの逐次水素化が妨げられ, シクロヘキセン選択性が向上したものと推定した.

The direct reaction of ethylene oxide with trioxane was thought impossible because of the weak basicity of trioxane. Thus, as the initiation mechanism of the copolymerization of trioxane and ethylene oxide, the formation mechanism of dioxolane from ethylene oxide and formaldehyde was proposed by several authors, and this mechanism was thought plausible for a long time.We found a new reaction, that is, the direct reaction of ethylene oxide with trioxane and isolated a novel cyclic formal. From this novel cyclic formal, 1, 3, 5-trioxepane was formed and then dioxolane was also generated. A new intermediate, or a novel cyclic compound, was separated using a micro-distillation apparatus. Its chemical structure was confirmed using 1H-NMR, 13C-NMR, Mass spectrum and elemental analysis. The new isolated intermediate or novel cyclic compound was 1, 3, 5, 7-tetraoxacyclononane, which identifies a new direct reaction between trioxane and ethylene oxide. This reaction identified the precise initiation mechanism of the copolymerization of trioxane and ethylene oxide.

Many background models have been proposed to adapt to "illumination changes" and "dynamic changes" such as swaying motion of tree branches. However, the problem of background maintenance in complex environment, where foreground objects pass in front of stationary objects which cease moving, is still far from being completely solved. To address this problem, we propose a framework for multi-layered background modeling, in which we conserve the background models for stationary objects hierarchically in addition to the one for the initial background. To realize this framework, we also propose a spatio-temporal background model based on the similarity in the intensity changes among pixels. Experimental results on complex scenes, such as a bus stop and an intersection, show that our proposed method can adapt to both appearances and disappearances of stationary objects thanks to the multi-layered background modeling framework.