Abstract: Hand-written characters which have been turned to electronic data are input by input means, and fields of induction on the retina of the character image are calculated by field of induction estimating means. By using the fields of induction on the retina generated by the plurality of characters thus calculated, character region of each character is determined by character segmentation means, and individual characters are segmented from an array of characters. The field of induction on the retina of the segmented character is deformed to be matched with a field of induction on the retina of a character prepared in advance as a dictionary, and based on the magnitude of strain generated at the time of this deformation, difference between the fields of induction on the retina of different characters is evaluated quantitatively, and character recognition is carried out based on this estimation.

Abstract: The invention relates to optical character recognition and to apparatus for representing characters in a manner analogous to the field which one would expect to be induced on the retina of the eye by a particular character. A character to be recognized is compared with the same type of representation for prototypical characters and the determination made, using an elastic model, of the amount of deformation required to change the field of induction of the scanned character into the shape of that of the prototypical character. Character recognition is achieved by determining which prototypical character requires minimum energy to deform into that of the scan character or vice versa.

Abstract: A plurality of data stored in a data base memory are displayed on a display portion at random, and when a retriever designates a desired data among the displayed data by a keyboard or a pointing device, a CPU stores the data in an auxiliary memory, estimates data which seems to be desirable for the retriever based on the stored data, and selects estimated data from the data base memory as well as data which are not determined to be desirable, and these are displayed simultaneously. Therefore, the scope of desirable data can be narrowed, well reflecting the intention of the retriever.

Abstract: A speech recognition method in which input speech signals are converted to digital signals and then time sequentially converted to cepstrum coefficients or logarithmic spectra. Dynamic spectrum time sequence is obtained by time frequency filtering of cepstrum coefficients, or masked spectrum time sequence is obtained by time frequency masking of the logarithmic vector time sequence. Based on the dynamic cepstrum time sequence or masked spectrum time sequence obtained in this manner, speech is recognized.

Abstract: In an apparatus for examining gaze shift in depth direction, light emitting diodes B and E, or D and C are lit at a plurality of positions with different distance in depth direction from the subject, and eye movement of the subject at that time is detected by an eye movement detecting portion. The signal of detection is processed in a signal processing circuit and is applied to a calculating portion. The calculating portion determines whether or not there is a disorder in gaze shift in depth direction of the subject by using any or all of latency in eye movement, time constant, amplitude, change in convergence angle, number of occurrence of saccade, and asymmetry of left and right eye movements, in response to the detected output of the eye movement.

Abstract: A target is presented on a board for calculation, eye movement of a subject while he is gazing at the target is detected by an eye movement detecting portion, the output of which is converted into a digital signal in a signal processing circuit and it is applied to a calculating portion. The calculating portion calculates small involuntary movement data in response to the input digital signal to find a value of a fractal dimension and time dependence of the fractal dimension to determine a disease related to brain function, and the output of determination is provided from an output portion such as a printer.

Abstract: When an image is entered from an input means, a line extracting means extracts and smoothes the contour of that image to obtain a tangent line passing through a certain point on that contour. A region with the certain point as the origin is determined, where Hough transform of a straight line is carried out with the center point of that region as the origin. Hough transform is carried out for all the points adjacent to the origin to obtain an envelope. A tangent line is predicted by the obtained envelope.

Abstract: An image at a distant point and an image at a close point for generating depth perception of a subject are displayed on an image display monitor, and eye movement of the subject at that time is detected by an eye movement detecting portion. A calculating portion removes saccade component by using velocity or acceleration of the detected eye movement and calculates amplitude of vergence eye movement, change in convergence angle and cross-correlation of velocity or acceleration of the left and right eye movement to determine depth perception of the subject.

Abstract: When eye movement of a subject is detected by an eye movement detecting portion and head movement of the subject is detected by a head movement detecting portion, a line-of-sight calculating portion calculates the line-of-sight according to the detected outputs therefrom. A portion for generating an image to be presented generates an image as a task and presents the image to a task presenting portion. When a fixation point of the line-of-sight of the subject is determined by a fixation point determining portion, a mask image generated by a masking image generating portion is inserted to the task image by an image synthesizing portion, and it is presented at the task presenting portion. A portion for determining degree of advance of dementia determines degree of advance and possibility of Alzheimer's disease based on the particular movement of the line-of-sight of the subject at that time and on the percentage of correct answer to a prescribed task.

Abstract: Disclosed in a pattern recognition apparatus including a neural net including a plurality of mutually coupled excitatory element-inhibitory element paris. The elements in the neural net are coupled through pair coupling coefficients and an excitatory element coupling coefficient. A teacher signal generator outputs a successive teacher signal Q (t) and its infinitesimal variation .DELTA.Q (t) in accordance with a learning pattern, to apply the output teacher signal and the output small variation to a coupling coefficient controller. The coupling coefficient controller carries out a learning processing in accordance with the teacher signal, so that the coupling coefficients in the neural net are updated. An operation time necessary for a pattern recognition is shortened. In addition, since the learning processing is carried out by only a forward processing on time base, a learning time is also decrease.

Abstract: In a medical diagnostic apparatus utilizing line-of-sight detection, targets are successively presented from different angles to the subject by a target presenting apparatus, eye movement of the subject while the subject is gazing at the target is detected by an eye movement detecting portion, the head movement of the subject is detected by a head movement detecting portion, and in response to the detection outputs, the spatial movement of the line-of-sight, rotation velocity component and translational velocity component of the head movement as well as the velocity component of the eye movement are calculated by a calculating portion, and the results are output to an output portion.

Abstract: In the medical diagnostic apparatus employing line-of-sight detection, eye movement and head movement of a subject are detected by eye movement detecting portion and head movement detecting portion, respectively, and images of the field of view of the subject are picked up by a TV camera. A calculating portion calculates a locus of the line-of-sight, a locus of the eye movement, the locus of the head movement, and vectors of spatial movement of the line-of-sight in response to the detected eye movement and the head movement. An image processing portion reduces the picked up image, a superimposer superimposes the images of the loci, the vector images and the images of the reduced field of view on each other, and the resulting images are displayed on a CRT display.

Abstract: Flicks in one or both of eyes are measured by a sensor (1) employing a limbus reflection method. An output of the sensor is applied to and digitized by an A/D converter (2). The digitized output is applied to a computer (3). The computer calculates flicks data of eyeballs, then calculates fractal dimension simultaneously or individually in a time-elapsed manner or for a definite time period, and quantifies characteristics of the eyeball control systems from the value of the fractal dimension or the time dependence of the fractal dimension.

Abstract: A movement trajectory generating system of a dynamical system uses neural network units (1, 2, 3) including cascade connection of a first layer (11, 21, 31), a second layer (12, 22, 32), a third layer (13, 23, 33) and a fourth layer (14, 24, 34), to learn a vector field of differential equations indicating forward dynamics of a controlled object (4). Conditions concerning trajectories of a final point and a via-point of movement of the controlled object and locations of obstacles are given from a motor center (5). While smoothness of movement is ensured by couplings of electric synapses using errors with respect to those conditions as total energy, least dissipation of energy is attained, whereby trajectory formation and control input for realizing the trajectory are obtained simultaneously.