Kyushu University Academic Staff Educational and Research Activities Database
List of Presentations
Kenji Hara Last modified date:2021.08.17

Professor / Department of Visual Communication Design / Department of Communication Design Science / Faculty of Design


Presentations
1. Kenji Hara, Kohei Inoue, HDR Image Saliency Estimation by Convex Optimization, 27th IEEE International Conference on Image Processing, ICIP 2020, 2020.10.
2. Wei Zhang, Kohei Inoue, Kenji Hara, An Equivalence Between Log-Sum-Exp Approximation and Entropy Regularization in K-Means Clustering, IEICE NOLTA, 2019.12.
3. Kenji Hara, Kohei Inoue, Kiichi Urahama, Full-Reference Metric Adaptive Image Denoising, 26th IEEE International Conference on Image Processing, ICIP 2019, 2019.09.
4. Daiki Okazaki, An-shui Yu, Kenji Hara, Omnidirectional Saliency Map Generation by Yin-Yang Grid Method, ACM International Conference on Video and Image Processing 2018, 2018.12.
5. An Shui Yu, Hara Kenji, Kohei Inoue, Kiichi Urahama, Foreground Enlargement of Omnidirectional Images by Spherical Trigonometry, 24th International Conference on Pattern Recognition, ICPR 2018, 2018.11.
6. Shuoyan Zhang, Kohei Inoue, Kenji Hara, Kiichi Urahama, An Iterative Raster Scan Algorithm for Superpixel Segmentation, ITC-CSCC, 2018.07.
7. Kohei Inoue, Kenji Hara, Kiichi Urahama, Image Size-Preserving Visual Cryptography by Error Diffusion, ITC-CSCC, 2018.07.
8. Zixu Zhang, Kohei Inoue, Kenji Hara, Kiichi Urahama, Voxel-Artistic Rendering of Color Images, ITC-CSCC, 2018.07.
9. An Shui Yu, Hara Kenji, Kohei Inoue, Kiichi Urahama, Corner detection in fisheye images by modified Yin-Yang grid, 43rd Annual Conference of the IEEE Industrial Electronics Society, IECON 2017, 2017.12, We propose a framework to extend corner feature detection in standard rectangular images with less distortion to distorted circular images captured with fisheye lenses. To solve two problems of nonuniformity of spatial resolution and spherical polar coordinates singularity, our approach makes use of a modification in the Yin-Yang grid, which is an overset grid consisting of two latitude/longitude coordinate systems. The main contribution is to keep the use of existing corner detection programs. Experimental results on synthetic and real images demonstrate the effectiveness of our method..
10. Genki Ono, Kohei Inoue, Hara Kenji, Kiichi Urahama, Reversible data hiding using maximum and minimum filters for image interpolation, 6th IEEE Global Conference on Consumer Electronics, GCCE 2017, 2017.12, We propose a reversible data hiding method for embedding secret data in a scaled-up image of a cover image, where the scaled-up image is computed using maximum and minimum filters to achieve higher payload than that of conventional methods based on neighbor mean interpolation. Experimental results show that the proposed method outperforms conventional methods in both data-hiding capacity and image quality..
11. An-shui Yu, Kenji Hara, Kohei Inoue, kiichi urahama, Corner Detection in Fisheye Images by Modified Yin-Yang Grid, IECON 2017, 2017.10.
12. Genki Ono, Kohei Inoue, Kenji Hara, Kiichi Urahama, Reversible Data Hiding Using Maximum and Minimum Filters for Image Interpolation, GCCE, 2017.10.
13. An-shui Yu, Kenji Hara, Kohei Inoue, kiichi urahama, Foreground Enlargement on Spherical Images Based on Analytical Mechanics, IWAIT 2017, 2017.01.
14. Kohei Inoue, Hara Kenji, Kiichi Urahama, Reflectance spectra recovery with non-negativity constraints, 2016 International Symposium on Intelligent Signal Processing and Communication Systems, ISPACS 2016, 2017.01, We propose two methods for recovering the reflectance spectra of given colorimetric data by using the nonnegative constraints in reflectance spectra. We formulate the problem of reflectance spectra recovery as a non-negative least squares problem and solve it with two iterative methods. Experimental results demonstrate that the two methods give similar recovery results, where Macbeth ColorChecker data are used for recovering the reflectance spectra of Neugebauer primary colors. We also transform the recovered reflectance spectra into tristimulus values to visualize them, where an ad hoc scaling operation is introduced for brightening the recovered colors..
15. Zixu Zhang, Kohei Inoue, Kenji Hara, kiichi urahama, Image Downscaling Based on Neugebauer Model, ICISIP, 2016.09.
16. Hengjun Yu, Kohei Inoue, Kenji Hara, kiichi urahama, Color Error Diffusion Based on Neugebauer Model, ICISIP, 2016.09.
17. An-shui Yu, Kenji Hara, Kohei Inoue, kiichi urahama, A Corner Detection Approach for Fisheye Images, IWAIT 2016, 2016.01.
18. Hengjun Yu, Kohei Inoue, Kenji Hara, kiichi urahama, Piecewise Constant Histogram Specification for False Contour-Free Contrast Enhancement, ICISIP, 2015.09.
19. Kohei Inoue, Kenji Hara, kiichi urahama, A Linear-Time Method for Multi-Exposure Image Fusion, ICAI, 2015.06.
20. An-shui Yu, Kenji Hara, Kohei Inoue, kiichi urahama, Spatially-Variable Laplacian Edge Detector for Fisheye Images, IWAIT and IFMIA, 2015.01.
21. Kohei Inoue, Kenji Hara, kiichi urahama, Integral Image-Based Error Diffusion, PCS, 2013.12.
22. Kohei Inoue, Hara Kenji, Kiichi Urahama, A hybrid method for high density salt-and-pepper noise removal, 2013 9th International Conference on Signal-Image Technology and Internet-Based Systems, SITIS 2013, 2013.12, Recently, two methods for removing high density salt-and-pepper noise in images have been proposed by Esakkirajan et al. and Hong et al. However, their methods have not been yet compared experimentally. In this paper, we compare them and show some experimental results. Additionally, we propose a hybrid method which is derived by combining the two methods, and experimentally show that the proposed method achieves higher values of both peak signal-to-noise ratio (PSNR) and image enhancement factor (IEF) than the two methods..
23. Zihan YU, Kohei Inoue, Hara Kenji, Naoki Ono, Kiichi Urahama, Bilateral Minimum Filter for Dehazing Images, 2013.06.
24. Kohei Inoue, Hara Kenji, Kiichi Urahama, Integral image-based error diffusion, 2013 Picture Coding Symposium, PCS 2013, 2013.01, We propose a new error diffusion method based on integral images. The proposed method has only 3 error diffusion coefficients, and therefore is computationally efficient compared to the conventional error diffusion methods. Experimental results show that the proposed method is fast while keeping the halftone quality..
25. Kohei Inoue, Hara Kenji, Kiichi Urahama, Memory-efficient computation of high-dimensional integral images, 2013 2nd IAPR Asian Conference on Pattern Recognition, ACPR 2013, 2013.01, We propose a memory-efficient method for computing high-dimensional integral images. The proposed method generates exactly the same integral images as that generated by the conventional method, while using less memory. Experimental results show the effectiveness of the proposed method..
26. Kohei Inoue, Hara Kenji, Kiichi Urahama, A unified view of two-dimensional principal component analyses, Joint IAPR International Workshops on Structural and Syntactic PatternRecognition, SSPR 2012 and Statistical Techniques in Pattern Recognition,SPR 2012, 2012.11, Recently, two-dimensional principal component analysis (2D-PCA) and its variants have been proposed by several researchers. In this paper, we summarize their 2DPCA variants, show some equivalence among them, and present a unified view in which the non-iterative 2DPCA variants are interpreted as the non-iterative approximate algorithms for the iterative 2DPCA variants, i.e., the non-iterative 2DPCA variants are derived as the first iterations of the iterative algorithm started from different initial settings. Then we classify the non-iterative 2DPCA variants on the basis of their algorithmic patterns and propose a new non-iterative 2DPCA algorithm based on the classification. The effectiveness of the proposed algorithm is experimentally demonstrated on three publicly accessible face image databases..
27. Kohei Inoue, Hara Kenji, Kiichi Urahama, Symmetric generalized low rank approximations of matrices, 2012 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2012, 2012, Recently, the generalized low rank approximations of matrices (GLRAM) have been proposed for dimensionality reduction of matrices such as images. However, in GLRAM, it is necessary for users to specify the numbers of rows and columns in low rank matrices. In this paper, we propose a method for determining them semiautomatically by symmetrizing GLRAM. Experimental results show that the proposed method can determine the optimal ranks of matrices while achieving competitive approximation performance..
28. Kohei Inoue, Hara Kenji, Kiichi Urahama, Image and video clipping by weighted histogram intersection minimization, 2010 IEEE Region 10 Conference, TENCON 2010, 2010.12, We propose a method for clipping a rectangular region from an image by minimizing a weighted intersection of two color histograms which are constructed with pixels included in the inside and the outside of the rectangular region. Furthermore, we extend the clipping method for images to that for videos. Experimental results show that the proposed method can clip the regions of objects from images and remove the regions of backgrounds..
29. Kohei Inoue, Hara Kenji, Kiichi Urahama, Image resizing by row/column removal and addition based on hierarchical clustering, 2010 IEEE Region 10 Conference, TENCON 2010, 2010.12, We propose a method for removing rows and columns in an image by clustering the sets of rows and columns separately. We also propose a method for adding rows and columns to an image on the basis of the proposed row/column removal method. Experimental results show that the proposed methods can reduce or enlarge the sizes of given images, while the aspect ratios of the main objects in the images are preserved..
30. Kohei Inoue, Hara Kenji, Kiichi Urahama, Robust simultaneous low rank approximation of tensors, 3rd Pacific Rim Symposium on Image and Video Technology, PSIVT 2009, 2009.02, We propose simultaneous low rank approximation of tensors (SLRAT) for the dimensionality reduction of tensors and modify it to the robust one, i.e., the robust SLRAT. For both the SLRAT and the robust SLRAT, we propose iterative algorithms for solving them. It is experimentally shown that the robust SLRAT achieves lower reconstruction error than the SLRAT when a dataset contains noise data. We also propose a method for classifying sets of tensors and call it the subspace matching, where both training data and testing data are represented by their subspaces, and each testing datum is classified on the basis of the similarity between subspaces. It is experimentally verified that the robust SLRAT achieves higher recognition rate than the SLRAT when the testing data contain noise data..
31. Hara Kenji, Ko Nishino, Illumination and spatially varying specular reflectance from a single view, 2009 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 2009.01, Estimating the illumination and the reflectance properties of an object surface from a sparse set of images is an important but inherently ill-posed problem. The problem becomes even harder if we wish to account for the spatial variation of material properties on the surface. In this paper, we derive a novel method for estimating the spatially varying specular reflectance properties, of a surface of known geometry, as well as the illumination distribution from a specular-only image, for instance, captured using polarization to separate reflection components. Unlike previous work, we do not assume the illumination to be a single point light source. We model specular reflection with a spherical statistical distribution and encode the spatial variation with radial basis functions of its parameters. This allows us to formulate the simultaneous estimation of spatially varying specular reflectance and illumination as a sound probabilistic inference problem, in particular, using Csisźar's I-divergence measure. To solve it, we derive an iterative algorithm similar to expectation maximization. We demonstrate the effectiveness of the method on synthetic and real-world scenes..
32. Kohei Inoue, Hara Kenji, Kiichi Urahama, Robust multilinear principal component analysis, 12th International Conference on Computer Vision, ICCV 2009, 2009, We propose two methods for robustifying multilinear principal component analysis (MPCA) which is an extension of the conventional PCA for reducing the dimensions of vectors to higher-order tensors. For two kinds of outliers, i.e., sample outliers and intra-sample outliers, we derive iterative algorithms on the basis of the Lagrange multipliers. We also demonstrate that the proposed methods outperform the original MPCA when datasets contain such outliers experimentally..
33. Hara Kenji, Kohei Inoue, Kiichi Urahama, Separation of layers from images containing multiple reflections and transparency using cyclic permutation, 2009 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2009, 2009, In the paper, we propose a new method for blind separation of an arbitrary number of images from a set of their linear mixtures with unknown coefficients. This approach is as follows. We first introduce a novel multiple correlation between one image and a set of multiple images. Then this multiple correlation leads us to provide a set of simultaneous linear equations for updating each mixture of images. Finally, source images are recovered by iterating between solving the sets of equations and cyclically permuting the mixtures of images. The technique can be applied for extracting multiple layers from images containing multiple reflections and transparency..
34. Yumi Iwashita, Ryo Kurazume, Hara Kenji, Seiichi Uchida, Kenichi Morooka, Tsutomu Hasegawa, Fast 3D reconstruction of human shape and motion tracking by parallel fast level set method, 2008 IEEE International Conference on Robotics and Automation, ICRA 2008, 2008.09, This paper presents a parallel algorithm of the Level Set Method named the Parallel Fast Level Set Method, and its application for real-time 3D reconstruction of human shape and motion. The Fast Level Set Method is an efficient implementation algorithm of the Level Set Method and has been applied to several applications such as object tracking in video images and 3D shape reconstruction using multiple stereo cameras. In this paper, we implement the Fast Level Set Method on a PC cluster and develop a real-time motion capture system for arbitrary viewpoint image synthesis. To obtain high performance on a PC cluster, efficient load-balancing and resource allocation algorithms are crucial problems. We develop a novel optimization technique of load distribution based on the estimation of moving direction of object boundaries. In this technique, the boundary motion is estimated in the framework of the Fast Level Set Method, and the optimum load distribution is predicted and performed according to the estimated boundary motion and the current load balance. Experiments of human shape reconstruction and arbitrary viewpoint image synthesis using the proposed system are successfully carried out..
35. Kenichi Morooka, Xian Chen, Ryo Kurazume, Seiichi Uchida, Hara Kenji, Yumi Iwashita, Makoto Hashizume, Real-time nonlinear FEM with neural network for simulating soft organ model deformation, 11th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2008, 2008, This paper presents a new method for simulating the deformation of organ models by using a neural network. The proposed method is based on the idea proposed by Chen et al. [2] that a deformed model can be estimated from the superposition of basic deformation modes. The neural network finds a relationship between external forces and the models deformed by the forces. The experimental results show that the trained network can achieve a real-time simulation while keeping the acceptable accuracy compared with the nonlinear FEM computation..
36. Hara Kenji, Yuuki Kabashima, Yumi Iwashita, Ryo Kurazume, Tsutomu Hasegawa, Robust 2D-3D alignment based on geometrical consistency, 6th International Conference on 3-D Digital Imaging and Modeling, 3DIM 2007, 2007.12, This paper presents a new registration algorithm of a 2D image and a 3D geometrical model, which is robust for initial registration errors, for reconstructing a realistic 3D model of indoor scene settings. One of the typical techniques of pose estimation of a 3D model in a 2D image is the method based on the correspondences between 2D photometrical edges and 3D geometrical edges projected on the 2D image. However, for indoor settings, features extracted on the 2D image and jump edges of the geometrical model, which can be extracted robustly, are limited. Therefore, it is difficult to find corresponding edges between the 2D image and the 3D model correctly. For this reason, in most cases, the relative position has to be manually set close to correct position beforehand. To overcome this problem, in the proposed method, firstly the relative pose is roughly estimated by utilizing geometrical consistencies of back-projected 2D photometrical edges on a 3D model. Next, the edge-based method is applied for the precise pose estimation after the above estimation procedure is converged. The performance of the proposed method is successfully demonstrated with some experiments using simulated models of indoor scene settings and actual environments measured by range and image sensors..
37. Hara Kenji, Ryo Kurazume, Kohei Inoue, Kiichi Urahama, Segmentation of images on polar coordinate meshes, 14th IEEE International Conference on Image Processing, ICIP 2007, 2007.12, The Chan-Vese level set algorithm has been successfully applied to segmentation of images on Cartesian coordinate meshes, including ordinary planar images. In this paper we present a Chan-Vese model for segmentation of images on polar coordinate meshes, such as topography and remote sensing images. The image segmentation is accomplished by formulating the associated evolution equation in the polar coordinate system and then numerically solving the partial differential equation on an overset grid system called the Yin-Yang grid, which is free from the problem of singularity at the poles. We include examples of segmentations of real earth data that demonstrate the performance of our method..
38. Hara Kenji, Kohei Inoue, Kiichi Urahama, Automated separation of reflections from a single image based on edge classification, 2nd International Conference on Computer Vision Theory and Applications, VISAPP 2007, 2007, Looking through a window, the object behind the window is often disturbed by a reflection of another object. In the paper, we present a new method for separating reflections from a single image. Most existing techniques require the programmer to create an image database or require the user to manually provide the position and layer information of feature points in the input image, and thus suffer from being extremely laborious. Our method is realized by classifying edges in the input image based on the belonging layer and formalizing the problem of decomposing the single image into two layer images as an optimization problem easier to solve based on this classification, and then solving this optimization with a pyramid structure and deterministic annealing. As a result, we are able to accomplish almost fully automated separation of reflections from a single image..
39. Daisuke Miyazaki, Hara Kenji, Katsushi Ikeuchi, Photometric Stereo beyond Glass: Active Separation of Transparent Layer and Five-Light Photometric Stereo with M-Estimator Using Laplace Distribution for a Virtual Museum, NINTH IEEE INTERNATIONAL CONFERENCE ON COMPUTER VISION, 2007.
40. Yumi Iwashita, Ryo Kurazume, Tsutomu Hasegawa, Hara Kenji, Robust motion capture system against target occlusion using Fast Level Set Method, 2006 IEEE International Conference on Robotics and Automation, ICRA 2006, 2006.12, This paper introduces a new motion capture system for recovering 3D models of multiple persons separately and robustly against occlusion in real-time. Various markerless motion capture systems using video cameras have been proposed so far. However, in case that there are multiple persons in the scene at the same time, it is quite difficult to reconstruct a precise 3D model of each person separately due to the occlusion between them. To deal with this problem, the Fast Level Set Method is utilized in the proposed system for integrating stereo range data which is captured by multiple stereo cameras located around the target people. To reconstruct precise 3D models in real-time, the proposed system is implemented on a PC cluster with seven PCs and four stereo cameras. Tracking experiment of multiple persons and real-time reconstruction of 3D human models using the proposed system are successfully carried out..
41. Hara Kenji, Ko Nishino, Katsushi Ikeuchi, Multiple light sources and reflectance property estimation based on a mixture of spherical distributions, Proceedings - 10th IEEE International Conference on Computer Vision, ICCV 2005, 2005.12, In this paper, we propose a new method for simultaneously estimating the illumination of the scene and the reflectance property of the object from a single image. We assume that the illumination consists of multiple point sources and the shape of the object is known. Unlike previous methods, we will recover not only the direction and intensity of the light sources, but also the number of light sources and the specular reflection parameter of the object. First, we represent the illumination on the surface of a unit sphere as a finite mixture of von Mises-Fisher distributions by deriving a spherical specular reflection model. Next, we estimate this mixture and the number of distributions. Finally, using this result as initial estimates, we refine the estimates using the original specular reflection model. We can use the results to render the object under novel lighting conditions..
42. Hara Kenji, Kohei Inoue, Kiichi Urahama, Adaptive image translation for painterly rendering, 9th IAPR Conference on Machine Vision Applications, MVA 2005, 2005, In the paper, we present a new method of converting a photo image to a synthesized painting image following the painting style of an example painting image.The proposed method uses a hierarchical and adaptive patch-based approach to both the synthesis of painting styles and preservation of scene details.This approach can be summarized as follows. The input photo image is represented as a set of patches divided adaptively using a distance transform technique. Then the mapping between the input photo and example painting images is efficiently inferred using Bayesian belief propagation recursively..
43. Kohei Inoue, Hara Kenji, Kiichi Urahama, Matrix principal component analysis for image compression and recognition, 9th IAPR Conference on Machine Vision Applications, MVA 2005, 2005.
44. Yumi Iwashita, Ryo Kurazume, Tokuo Tsuji, Tsutomu Hasegawa, Hara Kenji, Fast implementation of level set method and its real-time applications, 2004 IEEE International Conference on Systems, Man and Cybernetics, SMC 2004, 2004.12, The level set method (LSM) has been widely used for various applications such as motion tracking and 3D geometrical modeling. However, the calculation cost of reinitialization and updating of an implicit function is considerably expensive as compared with conventional active contour models such as "Snakes". To tackle this problem, we propose an ef cient algorithm of the LSM named the Fast Level Set Method(FLSM). This paper introduces some experiments based on the FLSM, including 2D real-time tracking of moving objects in video images, and 3D simultaneous motion capture system of multiple targets using stereo range images..
45. Daisuke Miyazaki, Robby T. Tan, Hara Kenji, Katsushi Ikeuchi, Polarization-based inverse rendering from a single view, NINTH IEEE INTERNATIONAL CONFERENCE ON COMPUTER VISION, 2003.12, This paper presents a method to estimate geometrical, photometrical, and environmental information of a singleviewed object in one integrated framework under fixed viewing position and fixed illumination direction. These three types of information are important to render a photorealistic image of a real object. Photometrical information represents the texture and the surface roughness of an object, while geometrical and environmental information represent the 3D sliape of an object and the illumination distribution, respectively. The proposed method estimates the 3D shape by computing the surface normal from polarization data, calculates the texture of the object from the diffuse only reflection component, determines the illumination directions from the position of the brightest intensity in the specular reflection component, and finally computes the surface roughness of the object by using the estimated illumination distribution..
46. Hara Kenji, Ko Nishino, Katsushi Ikeuchi, Determining reflectance and light position from a single image without distant illumination assumption, Proceedings: Ninth IEEE International Conference on Computer Vision, 2003, Several techniques have been developed for recovering reflectance properties of real surfaces under unknown illumination conditions. However, in most cases, those techniques assume that the light sources are located at infinity, which cannot be applied to, for example, photometric modeling of indoor environments. In this paper, we propose two methods to estimate the surface reflectance property of an object, as well as the position of a light source from a single image without the distant illumination assumption. Given a color image of an object with specular reflection as an input, the first method estimates the light source position by fitting to the Lambertian diffuse component, while separating the specular and diffuse components by using an iterative relaxation scheme. Moreover, we extend the above method by using a single specular image as an input, thus removing its constraints on the diffuse reflectance property and the number of light sources. This method simultaneously recovers the reflectance properties and the light source positions by optimizing the linearity of a log-transformed Torrance-Sparrow model. By estimating the object's reflectance property and the light source position, we can freely generate synthetic images of the target object under arbitrary source directions and source-surface distances..
47. Ko Nishino, Hara Kenji, Robby T. Tan, Daisuke Miyazaki, Katsushi Ikeuchi, Photometric Aspects on the Preservation of Cultural Assets, NINTH IEEE INTERNATIONAL CONFERENCE ON COMPUTER VISION, 2002.
48. F. Nagata, K. Watanabe, S. Hashino, H. Tanaka, T. Matsuyama, Hara Kenji, Polishing robot using a joystick controlled teaching system, 2000.01, An impedance model following force control, using a position/orientation compensator based on joystick taught data, is proposed for an industrial robot with an open architecture controller. The method is characterized by two requirements: 1) to polish an object with the desired contact force and orientation; and 2) no conventional complicated teaching process is required. The effectiveness and potential of the proposed method are demonstrated through some experiments concerning with a polishing task using the JS-10 industrial robot..
49. Hara Kenji, H. Zha, T. Hasegawa, Topology-adaptive modeling of objects by a level set method with multi-level stopping conditions, International Conference on Image Processing (ICIP'99), 1999, Level set methods were proposed mainly by mathematicians for extracting bounding contours of multiple regions or shapes with holes. In the paper, we propose a new method of constructing a surface/solid model of a 3-D object of arbitrary topology on the basis of the approach. Given sensor data covering the whole object surface, the method begins with an initial approximation of the object by evolving a closed-surface into a model topologically equivalent to the real object. The refined approximation is then performed by deforming it at an optimally tuned speed. By numerically solving the level set equations for tracking the evolving surfaces, it becomes possible to handle 3-D objects of arbitrary topology while maintaining high accuracy in the shape fitting..
50. Hara Kenji, H. Zha, T. Hasegawa, Regularization-Based 3-D Object Modeling from Multiple Range, International Conference on Image Processing (ICIP'99), 1998.