日浦 慎作

We evaluated the performance of our automated computerized scheme for determining the likelihood of malignancy of pulmonary nodules on high-resolution computed tomography (HRCT) and positron emission tomography (PET) scans. Our database contained 36 primary lung cancers and 9 benign nodules. After the nodule location was identified by a radiologist, the boundary surface of the nodule was segmented automatically using a spiral-scanning technique. Objective nodular features were assessed by quantitative analysis of the nodular shape and on gray-level histograms of the interior and exterior regions. The likelihood of malignancy was determined by a support vector machine. The performance of our technique in distinguishing between benign and malignant nodules was evaluated by receiver operating characteristic (ROC) analysis. The area under the ROC curve (AUC) value obtained by using HRCT features alone was 0.87. The integration of PET features into the AUC value resulted in a significant improvement from 0.87 to 0.94 (P< 0.01). The AUC value obtained from simultaneous selection of HRCT and PET features was 0.97. A statistically significant difference (P< 0.01) was observed between the result obtained by simultaneous HRCT and PET feature selection (AUC=0.97) and that by integration of PET features (AUC=0.94). Our automated computerized scheme for determining the likelihood of malignancy may help radiologists to differentiate between benign and malignant pulmonary nodules on HRCT and PET scans.

Calibration for calculating the external parameter of camera-projector system is important, to make high-precision 3-D measurement by 3-D endoscope using the grid active stereo. In this paper, we propose a new calibration method with a sphere calibration object in a geometrically flexible camera-projector system, and we examine the validity of the proposed method. As the result of comparing using conventional calibration with a plane calibration object, it is clarified that the proposed method resolves the scaling imprecision, that is a problem of the conventional method. Thus, the proposed method has higher potential of measurement accuracy. In addition, it is the advantage of the proposed method that precision estimation of calibration parameters with a low number of input images is possible.

The aim of this study is to improve the image quality using a post process rather than a correction process at acquisition time. We used a smoothing filter that is widely used on a compact digital camera. Especially for nuclear medicine, when we use a short acquisition time, we will get images that have a large increase in statistical noise. For those images, we validated the efficiency of the smoothing filter by assessing two characteristic parameters. In addition, we defined the best smoothing filter parameters to get stable images that reduced the influence of statistical noise.

本論文では撮像素子の各画素の受光感度分布をそれぞれランダムな形状に符号化することで,複数の画像を入力とする超解像処理を高性能化することができることを示す.既存の撮像素子の各画素の受光感度分布は方形であるため,画像の入力過程は原画像に対し方形を畳み込んだものをサンプリングしていることと等価である.このとき畳込みにより原画像の高空間周波数成分が失われることが超解像の性能を制限する要因となっている.そこで本研究では,撮像素子の上に細かい黒色粉末を散布することで,各画素の受光感度分布をランダムに符号化する.これに圧縮センシング分野で用いられているスパース正則化を組み合わせることで,少ない枚数の入力画像から高解像度な画像を復元し,既存の撮像素子を備えたカメラより超解像の性能が向上することを実機を用いた実験により示す.

ステレオ法はシーンに対して光を投影することなくシーンの奥行マップを得ることができる手法であるが,被写体上のテクスチャやエッジが繰返し模様であったりエピポーラ線と並行である場合に画像間の対応付けが不安定になるという問題がある.一方,錯乱円径が距離によって変化することを利用し距離を得るDepth from Defocusは,レンズの口径によって奥行精度が限定されるという欠点をもつ.そこで本研究では,ステレオカメラの左右のカメラに異なる合焦距離を与え,更に符号化開口を導入することでステレオ法とDepth from Defocusを融合し,両手法の利点を兼ね備えた安定度と精度をもつ奥行推定方法を提案する.

We propose a polarization-based method to enhance the visibility of an image by canceling the haze effect. Haze is a natural phenomenon that degrades the visibility of a scene. Aerosols in air reflect sunlight and cause polarization. Therefore, we analyze the polarization state of the observed light to remove the haze effect from a captured image. Our approach is to use two reference objects that are known a priori in estimating the parameters of the haze effect. Once the parameters are known, we can improve the image so that the scene is clearly visible. We also present an experimental result using a commercial polarization camera, which can obtain the polarization state of the scene. © 2013 IEEE.

This paper shows a random and distinct shape of each pixel improves the performance of super-resolution using multiple input images. Since the spatial light sensitivity distribution in each pixel of an image sensor is rectangular and identical, the process of imaging is equivalent to the point sampling of blurred image which is a result of convolution of a rectangle with the original image. The convolution results in a loss of the high spatial frequency component of the original image, which limits the performance of super-resolution. Thus, we sprayed a fine-grained black powder on an image sensor to give a random code to the spatial light sensitivity distribution in each pixel. This approach was combined with a reconstruction technique based on sparse regularization, which is commonly used in compressed sensing, in an experiment with an actual setup. A high-resolution image was reconstructed from a limited number of input images and the performance of super-resolution was significantly improved.

Metamerism is a phenomenon where two objects recognized as having different colors under one light are also recognized as having the same color under another light. This research proposes technology for actualizing artistic illusion that exploits metamerism. Specifically, the purpose of the research relates to automatic calculation of blending ratios of oil paints that cause metamerism to occur under specific light sources. We entails metamerism occurring between three types of object colors under two types of light sources. Also, we utilize plausible reflection model for the mixture of oil paints.

In this paper we propose a novel depth measurement method by fusing depth from defocus (DFD) and stereo. One of the problems of passive stereo method is the difficulty of finding correct correspondence between images when an object has a repetitive pattern or edges parallel to the epipolar line. On the other hand, the accuracy of DFD method is inherently limited by the effective diameter of the lens. Therefore, we propose the fusion of stereo method and DFD by giving different focus distances for left and right cameras of a stereo camera with coded apertures. Two types of depth cues, defocus and disparity, are naturally integrated by the magnification and phase shift of a single point spread function (PSF) per camera. In this paper we give the proof of the proportional relationship between the diameter of defocus and disparity which makes the calibration easy. We also show the outstanding performance of our method which has both advantages of two depth cues through simulation and actual experiments.

In this paper, we propose a novel 3-D endoscope system by using grid-based active stereo. In the proposed system, projection of a waved-grid pattern that consists of vertical and horizontal sinusoidal lines realizes accurate shape acquisition in sub-pixel accuracy. We develop a small pattern projector implementable to a head of a ready-made endoscope, and examine 3-D shape reconstruction by actual equipment. As the result of the measurement for a known-shaped object, which is a hexagonal cylinder, the error of length measurement is below 0.9% and the error of angle measurement is below 2.2%. We make a measurement of animal organ meat, and confirm that the system can reconstruct a 3-D shape of the organ surface. © 2013 IEEE.

We propose a method to observe cardiac beat from 3D shape information of body surface by using the active stereo with waved-grid pattern projection, and report preliminary experiments to evaluate validities of the proposed method. By comparing results of our method with those of electrocardiogram (ECG), we confirmed sufficient correspondences between peak intervals of depth changes between contiguous frames measured by the active stereo and R-R intervals measured by ECG. We proposed the visualization of the spatial distribution of depth change plotted on the 3D shape of chest surface. We confirm that the spatial phase difference, which is caused by heart pump ability, appears in the 3-D shape change of chest surface. © 2013 IEEE.

In this paper, we present an efficient method to remove shadow and specular components from multiple images taken under various lighting conditions. We call the task photometric linearization because Lambert's law of diffuse component follows linear subspace model. The conventional method[1] based on a random sampling framework make it possible to achieve the task, however, it contains two problems. The first is that the method requires manual selection of three images from input images, and the selection seriously affects to the quality of linearization result. The other is that an enormous number of trials takes a long time to find the correct answer. We therefore propose a novel algorithm using the PCA (principal component analysis) method with outlier exclusion. We used knowledge of photometric phenomena for the outlier detection and the experiments show that the method provides fast and precise linearization results. Additionally, as an application of the proposed method, we also present a method of lossless compression of HDR images taken under various lighting conditions. © Springer-Verlag Berlin Heidelberg 2013.

We propose a method to observe cardiac beat from 3D shape information of body surface by using grid-based active stereo, and report preliminary experiments to evaluate validities of the proposed method. By comparing results of our proposed with those of electrocardiogram (ECG), we confirmed sufficient correspondences between peak intervals of inter-frame depth changes measured by the active stereo and R-R intervals measured by ECG. We tried the visualization of the spatial distribution of inter-frame depth change plotted on the 3D shape of chest region. And, the shape change by cardiac beat is mainly found on the left side of the chest region. © 2013 Springer-Verlag.

We have estimated scattering properties of participating media using rendered images. We have rendered multiple scattering in participating media using photon mapping. Then, we estimate the scattering properties by comparing the generated images and real images.

This paper describes a method to determine the direction of a light source and the distribution of diffuse reflectance from two images under different lighting conditions. While most inverse-rendering methods require 3 or more images, we investigate the use of only two images. Using the relationships between albedo and light direction at 6 or more points, we firstly show that it is possible to simultaneously estimate both of these if the shape of the target object is given. Then we extend our method to handle a specular object and shadow effect by applying a robust estimation method. Thorough experimentation shows that our method is feasible and stable not only for well controlled indoor scenes, but also for an outdoor environment illuminated by sunlight.

We propose a new non-contact measurement of cardiac beat from 3D shape information of body surface by using grid-based active stereo, and basically examine the validity of the proposed method. By simultaneous measurement with our proposed method and ECG, there are sufficient correspondences between peak intervals of inter-frame depth changes measured by our method and R-R intervals measured by ECG.

The human brain recognizes pictures that are first obtained by the photosensitive cells in the retina. Color is a visual perception composed by the stimulus of three kinds of photoreceptors called L, M, and S cones. The human eye often recognizes different spectral distribution as same color since each cone has wide spectral response. This phenomenon is known as metamerism. Our research project aims to innovate a novel form of artistic illusion by fully making use of metamerism. This paper proposes a method which estimates the mixture ratio of paints that can cause metamerism. Copyright is held by the author / owner(s).

Polarization is physical phenomena of the light which gives rich information of the scene even though it is undetectable by human eyes. In this paper, we analyze the polarization state of the light reflected on black specular objects in order to estimate the surface normal of the objects. Since polarization image of single view is not enough to uniquely determine the surface normal, we observe the object from multiple views. A rough estimate of the object geometry is obtained a priori by space carving so that the polarization state of the reflected light can be analyzed at the same surface point under multiple views. Unlike the photometric stereo or the multiview stereo which cannot estimate the surface normal and the geometry of black specular objects, the proposed method which is based on the polarization analysis and the space carving can estimate the surface normal and the geometry of black specular objects.

Recently 3D scanning systems are capable of modeling entire dense shapes that evolve over time with a single scan (a. k. a. one-shot scan). In particular, structured-light-based systems have emerged as one-shot shape reconstruction methods that project a static grid pattern onto the object surface. This pattern allows the scanning of moving objects while still maintaining dense reconstruction. One-shot scanning systems are then capable of producing 3D+t (a. k. a. 4D) spatio-temporal models with millions of points. As a consequence, effective 4D geometry compression schemes are required to face the need to store or transmit the huge amount of data, in addition to classical static 3D data. In this paper, we propose a 4D spatiotemporal rate-distortion (RD) optimized point cloud encoder via a curve-based representation of the point cloud, particularly well-suited for one-shot scanning systems. The object surface is naturally sampled in a series of curves, due to the grid pattern. This motivates our choice to leverage a curve-based representation to remove the spatial and temporal correlation of the sampled point along the scanning directions through a competitive-based predictive encoder that includes different spatio-temporal prediction modes through an RD cost computation control. Experimental results show the significant gain obtained with the proposed method.[Figure not available: see fulltext.] © 2012 3D Display Research Center and Springer-Verlag Berlin Heidelberg.

Recently it is relatively easy to produce digital point sampled 3D geometric models. In sight of the increasing capability of 3D scanning systems to produce models with millions of points, compression efficiency is of paramount importance. In this paper, we propose a novel competition-based predictive method for single-rate compression of 3D models represented as point cloud. In particular we aim at 3D scanning methods based on grid pattern. The proposed method takes advantage of the pattern characteristic made of vertical and horizontal lines, by assuming that the object surface is sampled in curve of points. We then designed and implemented a predictive coder driven by this curve-based point representation. Novel prediction techniques are specifically designed for a curve-based cloud of points, and been competing between them to achieve high quality 3D reconstruction. Experimental results demonstrate the effectiveness of the proposed method. © 2011 IEEE.

Indirect reflection component degrades the preciseness of 3-D measurement with structured light projection. In this paper, we propose a method to suppress the indirect reflection components by spatial synchronous detection of structured light modulated with MLS (Maximum Length Sequence, M-sequence). Our method exploits two properties of indirect components; one is the high spatial frequency component which is attenuated through the scattering of projected light, and the other is the geometric constraint between projected light and its corresponding pixel of camera. Several experimental results of measuring translucent or concave objects show the advantage of our method.

With the increasing demands for highly detailed 3D data, dynamic scanning systems are capable of producing 3D+t (a.k.a. 4D) spatio-temporal models with millions of points recently. As a consequence, effective 4D geometry compression schemes are required to face the need to store/transmit the huge amount of data, in addition to classical static 3D data. In this paper, we propose a 4D spatio-temporal point cloud encoder via a curve-based representation of the point cloud, particularly well-suited for dynamic structured-light-based scanning systems, wherein a grid pattern is projected onto the surface object. The object surface is then naturally sampled in a series of curves, due to the grid pattern. This motivates our choice to leverage a curve-based representation to remove the spatial and temporal correlation of the sampled point along the scanning directions through a competitive-based predictive encoder that includes different spatio-ternporal prediction modes. Experimental results show the significant gain obtained with the proposed method.

This paper describes a method to determine the direction of a light source and the distribution of diffuse reflectance from two images under different lighting conditions. While most inverse-rendering methods require 3 or more images, we investigate the use of only two images. Using the relationships between albedo and light direction at 6 or more points, we firstly show that it is possible to simultaneously estimate both of these if the shape of the target object is given. Then we extend our method to handle a specular object and shadow effect by applying a robust estimation method. Thorough experimentation shows that our method is feasible and stable not only for well controlled indoor scenes, but also for an outdoor environment illuminated by sunlight. © 2011 Information Processing Society of Japan.

In this paper, we propose a method to accurately measure the shape of objects by suppressing indirect reflection such as interreflection or subsurface scattering. We use a modulation with M-sequence shifted along the line of the slit light to be accurately detected on the captured image in two ways. This method has two advantages; one is the characteristics of propagation of higher spatial frequency components and the other is geometric constraint between the projector and the camera. Prior to the measurement, epipolar constraint is obtained through calibration, and then the phase consistency is evaluated to suppress the interreflection. The value of cross-correlation is used to suppress the dilation of the light caused by the subsurface scattering.

We propose a novel wide angle imaging system inspired by compound eyes of animals. Instead of using a single lens, well compensated for aberration, we used a number of simple lenses to form a compound eye which produces practically distortion-free, uniform images with angular variation. The images formed by the multiple lenses are superposed on a single surface for increased light efficiency. We use GRIN (gradient refractive index) lenses to create sharply focused images without the artifacts seen when using reflection based methods for X-ray astronomy. We show the theoretical constraints for forming a blur-free image on the image sensor, and derive a continuum between 1:1 flat optics for document scanners and curved sensors focused at infinity. Finally, we show a practical application of the proposed optics in a beacon to measure the relative rotation angle between the light source and the camera with ID information. © 2010 Information Processing Society of Japan.

We show a new camera based interaction solution where an ordinary camera can detect small optical tags from a relatively large distance. Current optical tags, such as barcodes, must be read within a short range and the codes occupy valuable physical space on products. We present a new low-cost optical design so that the tags can be shrunk to 3 m m visible diameter, and unmodified ordinary cameras several meters away can be set up to decode the identity plus the relative distance and angle. The design exploits the bokeh effect of ordinary cameras lenses, which maps rays exiting from an out of focus scene point into a disk like blur on the camera sensor. This bokeh-code or Bokode is a barcode design with a simple lenslet over the pattern. We show that a code with 1 5 mu m features can be read using an off-the-shelf camera from distances of up to 2 meters. We use intelligent binary coding to estimate the relative distance and angle to the camera, and show potential for applications in augmented reality and motion capture. We analyze the constraints and performance of the optical system, and discuss several plausible application scenarios.

Exaggerated defocus cannot be achieved with an ordinary compact digital camera because of its tiny sensor size, so taking pictures that draw the attention of a viewer to the subject is hard. Many methods are available for controlling the focus and defocus of previously taken pictures. However, most of these methods require custom-built equipment such as a camera array to take pictures. Therefore, in this paper, we describe a method for creating images focused at any depth with an arbitrarily blurred background from a set of images taken by a handheld compact digital camera that is moved at random. Our method can produce various aesthetic blurs by changing the size, shape, or density of the blur kernel. In addition, we demonstrate the potential of our method through a subjective evaluation of blurred images created by our system.

We propose a novel camera setup in which both the lens and the sensor are perturbed during the exposure. We analyze the defocus effects produced by such a setup, and use it to demonstrate new methods for simulating a lens with a larger effective aperture size (i.e., shallower depth of field) and methods for achieving approximately depth-independent defocus blur size. We achieve exaggerated, programmable, and pleasing bokeh with relatively small aperture sizes such as those found on cell phone cameras. Destabilizing the standard alignment of the sensor and lens allows us to introduce programmable defocus effects and achieve greater flexibility in the image capture process. © 2009 IEEE.

We propose a novel wide angle imaging system inspired by compound eyes of animals. Instead of using a single lens, well compensated for aberration, we used a number of simple lenses to form a compound eye which produces practically distortion-free, uniform images with angular variation. The images formed by the multiple lenses are superposed on a single surface for increased light efficiency. We use GRIN (gradient refractive index) lenses to create sharply focused images without the artifacts seen when using reflection based methods for X-ray astronomy. We show the theoretical constraints for forming a blur-free image on the image sensor, and derive a continuum between 1 : 1 flat optics for document scanners and curved sensors focused at infinity. Finally, we show a practical application of the proposed optics in a beacon to measure the relative rotation angle between the light source and the camera with ID information. ©2009 IEEE.

Exaggerated defocus can not be created with an ordinary compact digital camera because of its tiny sensor size, so it is hard to take pictures that attract a viewer to the main subject. On the other hand, there are many methods for controlling focus and defocus of previously taken pictures. However most of these methods require purpose-built equipment such as a camera array to take pictures. Therefore, in this paper we propose a method to create images focused at any depth with arbitrarily blurred background from the set of images taken by a handheld compact digital camera moved randomly. Using our method, it is possible to produce various aesthetic blurs by changing the size, shape or density of the blur kernel. In addition, we confirm the potential of our method through a subjective evaluation of blurred images created by our system.

Exaggerated defocus cannot be achieved with an ordinary compact digital camera because of its tiny sensor size, so taking pictures that draw the attention of a viewer to the subject is hard. Many methods are available for controlling the focus and defocus of previously taken pictures. However, most of these methods require custom-built equipment such as a camera array to take pictures. Therefore, in this paper, we describe a method for creating images focused at any depth with an arbitrarily blurred background from a set of images taken by a handheld compact digital camera that is moved at random. Our method can produce various aesthetic blurs by changing the size, shape, or density of the blur kernel. In addition, we demonstrate the potential of our method through a subjective evaluation of blurred images created by our system.