The effects of Gaussian noise and burst noise as electromagnetic noise on TV picture degradation were studied. Three kinds of still picture and four kinds of moving picture were used in our opinion tests. The distrubance was subjectively evaluated. The tendency of the picture degradation against the every pictures was almost the same, but that of between still picture and moving picture was a few different from each other in the case where the power of burst noise was small.

View full abstract

A new criterion using spectral and spatial information is proposed to classify multi-spectral remote sensing images or textured images. The images are modeled with a hierarchical Markov Random Field(MRF) model, which consists of the observed intensity process and the hidden class label process. The class labels are estimated according to the maximum a posteriori(MAP) criterion, however, some reasonable approximations are done to reduce the computational expenses. Finally a stepwise classification algorithm is derived and simulation and experimental results are shown to verify the validity of this algorithm.

View full abstract

We have developed a temporal pattern recognition networks with newly developed quantizer neuron chip(ONC) and applied it to the object recognition system. One of the biggest issues of an object recognition is the recognition with rotation invariance under a fluctuating noisy environment. The shape of the object is converted to a series of angles as a function of the circumference of the shape (φ-s data) and can be treated as a series of temporal patterns. The networks consist of a Multi Functional Layered Network(MFLN) with QNC and a layer of neurons with self feedback (self feedback layer). The self feedback layer unifies the temporal recognition results of networks with QNC during a certain period defined by the time constant of self feedback and this function can realize the function of selective attention to certain areas of a series of temporal patterns. As a result, the system realizes rotation invariance in recognition and we obtained 100% recognition accuracy of 50 trials with fluctuating noise taken by CCD camera.

View full abstract

In the present paper, we propose a method to recognize Japanese finger spelling. Hand region Images are automatically cut off the RGB color image(512×512[pixels]) in order to apply this method to recognition of general sign language expressed by the upper half of a body in the future. Our proposing method extracts region of stretched fingers from images by erosion-diloation method, and from these regions extract fundamental features such as finger vector. And also the method detects the status of contacted and stretched fingers through extracting boundary between fingers using gradient vector, gradient intensity, and use them as features. Based on these features, we have constructed a decision tree and tried to recognize finger spellings.

View full abstract

This paper present a method for personal identification considering probabilities of measured values. First, 8 feature points, such as outer eye comers, inner eye comers, nostrils and tips of the mouth, were extracted, then 8 definite lines were calculated. On the assumption that the distribution of the deviation of each definite lines is Gaussian, probabilities of measured 9 definite lines were calculated. As a result, for 10 registered samples and 191 unknown samples of 41 persons, 97% recognition rates are obtained.

View full abstract

We describe a method for estimating camera parameters from image sequences for image synthesis. The method consists of first establishing correspondence and then, estimating the parameters by fitting the correspondence data to a transformation model based on perspective mapping model and 3-D rotation and zoom. operation model. We show by simulations and experiments that the proposed method successfully estimates focal length from image sequence and explains very well the induced motion field of images undergoing camera operation: 3-D rotation and zoom, and significantly outperforms conventional estimation methods, especially for wide-angled images.

View full abstract

A homotopy describes the transformation of one arbitrary curve into another that shares the same endpoints. In this paper, we propose a deformable cylinder model, based on homotopy, in which an arbitrary surface interpolated between two contours via a blending function is transformed into another surface over time. We then show how this homotopic deformation can be applied to the realistic animation of human faces in a virtual space teleconferencing system. Specifically, we show that facial expressions such as wrinkling of the forehead and opening and closing of the mouth can be synthesized and animated in real time through 3D homotopic deformations.

View full abstract