In this paper, we investigate how the positioning and size of input and reference patterns affect the use of multiple reference patterns in a classical Joint Transform Correlator (JTC), and we produce an optical system using a JTC with multiple reference patterns. This system is applied to feature extraction from handwritten alphabetical letters, using four linear elements aligned vertically, horizontally, and obliquely at angles of ±45 degrees as features. The extracted features were input into a three-layer neural network to decide which letter was input. This system achieved an average recognition rate of 88.8% for 30 sets of handwritten alphabetical letters.

抄録全体を表示

We propose a highly efficient arithmetic coding method for bi-level images. In this paper, our main interest lies in generating bi-level images suitable for arithmetic coding. Hence, a bi-level quantizer is integrated in our coder. Bi-level images are generated by the bi-level quantizer where restrictions are imposed to improve coding efficiency under less degradation of image quality. These restrictions are provided by a priori knowledge on geometrical and statistical properties. We describe our arithmetic coding method and show simulation results for bi-level images including characters, figures and pseudo gray-scale images. Although our coding method is a semi-reversible coding in that only slight distortion is introduced, our simulation results show that the coder can reduce the total amount of code by 6.6%-27.6% more than conventional arithmetic coders.

抄録全体を表示

A method of extracting routes of walkers based on investigation of speed and direction change over a short period is proposed. Images in time series are taken alternately with two values of intervals. The position of a walker after a longer interval is determined by using a vector of movement obtained from images over a shorter interval. Good accuracy is compatible with short computation time. By using a trapezoidal mask, a portion of one person is extracted from a figure in which some walkers are connected. This mask is determined from the parallax of two cameras. Our method does not require identification of a person and is expected to decrease computation time. Routes of walkers were almost always extracted in our experiment.

抄録全体を表示

This study proposed one-dimensional stochastic diffusion models for spontaneous eye blink and analyzed burst patterns of the blinks during low vigilance. The models presumed the interblink interval distribution to be first-passage-time probability densities of the Ornstein-Uhlenbeck process.To estimate the parameters, the behavior of upper eyelids was measured when the vigilance of the subjects was high and low. As the results, the observed data was in good agreement with the model output and the threshold of a virtual blink generator changed with vigilance. It became evident that the Ornstein-Uhlenbeck first-passage-time models are valid spontaneous blink models and that the burst patterns occur because of a shift of the threshold.

抄録全体を表示

大画面における薄型・軽量なCRTとして考案された背面変調方式フラットCRTの問題点である線状カソードの振動を解決できるアーチカソードフラットCRTを考案し, コンピュータシミュレーションと実験によりディスプレイとしての有用性を確認した.

抄録全体を表示

スクリーン印刷法によるAC-PDPのMgO保護膜を検討した.焼成によりMgOになる液体バインダの添加により5000時間経過した後も輝度変化は40%以下であり, バインダを含まない場合の輝度半減期が800時間に比し, 長寿命化された.また, 放電維持電圧の変動も抑制されている.

抄録全体を表示