Ryo Takamatsu

A new operator for image structure analysis

In this paper, we propose a new operator called shape operator for describing image structure. We consider an image function as a surface, then describe a shape of each pixel comparing with its neighbourhood in terms of topographical shapes such as hill, dale, ridge, valley, etc. The shape operator is established by utilizing the eigenvalues of Hessian of an image function. We show how to derive this operator, and its interesting properties. Finally, we illustrate by examples how the shape operator can give the same interpretation of an image although the image is corrupted by shading effects, and how it can give the same interpretation of an image viewed in different viewpoints.

A remote collaboration system using hand-force virtual environment

Collaboration is a communication configuration which is a basic to human society. If a virtual working environment is connected by networking in order to allow sharing by multiple users, an efficient remote collaboration system will be realized. This paper presents a remote collaboration system which enables operators at remote sites to collaborate by maintaining a virtual environment that includes hand-force interaction over a network. To support smooth remote collaboration, the virtual environments at different sites must all be maintained the same. It is difficult, however, to realize simultaneous access of multiple users to the same object. It is shown in this paper that the hand-force virtual environment enables multiple users to have simultaneous access to the same object, so that smooth remote collaboration is supported. It is also shown that the sense of shared activity is enhanced by hand-force environment.

Image classification using the surface-shape operator and multiscale features

This paper introduces new features for describing image patterns. We integrate the concepts of multiscale image analysis, aura matrix to define image features, and to obtain the features having robustness with illumination variations and shading effects, we analyse images based on the topographic structure described by the surface-shape operator. Then, illustrate usefulness of the proposed features with texture classifications. Results show that the proposed features extracted from multiscale images work much better than those from a single scale image, and confirm that the proposed features have robustness with illumination and shading variations. By comparisons with the multiresolution simultaneous autoregressive features using Mahalanobis distance and Euclidean distance, the proposed multiscale features give better performances for classifying the entire Brodatz textures (1966).

A new operator for describing topographical image structure

To develop a computer vision system, it is necessary to define a proper image structure representation to be used for interpreting images efficiently. In this paper, we propose a new operator called shape operator for describing topographical image structure. We consider an image function as a surface, then describe a shape of each pixel comparing with its neighbourhood in terms of topographical shapes such as hill, dale, ridge, valley, etc. The shape operator is established by utilizing the eigenvalue of Hessian of an image function. It is expressed in an explicit form in terms of the second order partial derivatives of an image function. We show how to derive this operator, its interesting properties, and an application for texture classification. Experimental results show its good performance for discriminating texture imaged. Especially, it can give the same interpretation of an image reflected in different lighting conditions since it has the invariance properties under linear and monotonic gray tone transformations.

A hierarchical description of digital grayscale images based on image dipoles

Deals with a hierarchical description of digital grayscale images based on non-local structural features, called image dipoles. An image dipole is defined to be a region whose lines of slope emanate from the same maximum, and reach the same minimum. Incorporating an image dipole based representation of an image into a scale-space filtering, we propose in this paper a hierarchical image dipole description of the image. The proposed description consists of image dipole networks of some scales, and linking of image dipoles across the scales such that the description contains all local transformations of the image dipole network in the scale-space.

New operator for determining surface shapes

In this paper, we propose a new operator called the Surface- Shape operator (SS-operator) for determining local surface shape, and use it for analyzing image structure. We consider an image function as a bivariate surface, then describe a shape of each pixel comparing with its neighborhood. In a geometry viewpoint, types of surface shapes are elliptic, hyperbolic, parabolic, etc. Here, we label them based on topographical structure and used a statistic approach to measure spatial distribution of these shapes overall the surface. The Surface-Shape operator is established by utilizing the eigenvalues of Hessian of an image function. We show how to derive this operator and its interesting properties, and give examples to demonstrate its usefulness in practical uses. Finally, we illustrate its roles for describing texture images by using co-occurrence matrices, a statistical measure, to represent its numerical description, and show its good performance for suppressing shading effects as well.

Understanding of Ridge-Valley Lines on Image -Intensity Surfaces in Scale-Space

When a grey-scale image is regarded as a surface, the considerable topographic primal structures are stationary points, ridge and valley lines. The ridge and valley lines are rich in information because they pass through all stationary points, and are invariant with respect to monotonic image transformations. As images can be interpreted at multiresolutions by using scale-space filtering, this paper presents an analysis of the local topology of ridge-valley lines and its change in the scale-space. The result shows that the topology of ridge-valley lines along a stationary path will change when the path reaches a singular stationary point or an umbilicus; whereas the change at any non-stationary point occurs when a ridge-valley line gets contact with another or vanishes.

Space browsing interface based on head posture information

We propose an interface to browse 3D virtual space in a human scale virtual environment using a scrolling technique. Our method is based on head posture estimation by processing the image of an oval marker. Because the marker is attached to a cap, the user can wear the marker easily and naturally. Our method feels natural because the direction of scrolling is consistent with the users direction of attention. The estimation of the position and direction of the marker is based on image moment features which makes the system robust to noise and allows real-time operation without any special hardware.

The surface-shape operator and its application to image retrieval

This paper proposes an operator called the surface-shape sperator for describing image structures. By considering an image function as a surface, the surface-shape operator is established for describing the shape of each pixel compared with its neighbourhood, In this paper,the surface-shape operator is used as a pre-processing for describing the multiscale topographic structures of images in scale-space representation. Then, image features are extracted from the transformed images instead of the original ones. The surface-shape operator has invariant properties under linear and monotonic gray tone transformations, and is insensitive to additive noises modelled by linear functions, so it has robustness with brightness variations and shading effects. Finally, we show the usefulness of the proposed approach with the image retrieval application. By comparing the Wold features, the multiresolution simultaneous autoregressive features and the proposed features, the proposed features can provide the best pattern retrieval accuracy.

Viewpoint Estimation based on Moment Feature.

This paper proposes viewpoint estimation method intended to be used as the basic device to estimate users intension in a visual environment. Our method is based on moment feature which represents non-local feature of image, such as the gravity center or the breadth, and has less computational cost and considerable robustness to noise. We also propose the form and use of marker suitable for our purpose. That is, putting an oval marker on the top of a users head. Experimental result shows that our method has good performance.