摘要:
An imaging apparatus wherein a captured image is subjected to a rotational transform so as to correct the inclination of one or more subjects in the captured image. The imaging apparatus defines, with reference to a physical inclination detected by a sensor, an angle range in which the image is to be processed. The imaging apparatus then processes the captured image and detects lines therein and inclination angles thereof. The imaging apparatus then creates histograms indicative of frequency distributions of the inclination angles of those lines, and selects, in the created histograms, an inclination angle having a frequency that satisfies a predetermined criterion, as a rotational angle to be used for the rotational transform.
摘要:
A wearable camera that is an imaging device includes a sensing unit that senses motion of a camera, and a prediction unit that calculates a pixel shift distance (d) between frames based on the results of the sensing. The prediction unit obtains a position (p1′) by moving a feature point (p1) extracted from the previous frame image by the pixel shift distance (d), and the setting unit sets a window (w4) having a position (p1′) as the center in the current frame image (Ft+1). A search unit searches the window (w4) and detects a feature point (q1) corresponding to the feature point (p1).
摘要:
A picture classifying unit in a video scene classifying device classifies obtained plural pictures into plural clusters based on a classification index which indicates a feature of the pictures. A cluster selecting unit selects at least one cluster among the classified clusters. An index generating unit assigns the index to at least one picture among the pictures which constitute the selected cluster.
摘要:
To provide an image processing device that calculates a parameter used for correcting large video jitter with high accuracy even when the accuracy of a sensor for measuring a movement of a photographing device is low. The image processing device includes: a constraint condition generating unit (130) that generates a constraint condition using sensor information such that a parameter value falls within a range; and a parameter calculating unit (140) that calculates the parameter according to the constraint condition. The constraint condition generating unit (130) has: a feature vector generating unit (131) that generates a feature vector showing features of the sensor information; and a motion classifying unit (such as 132) that identifies a movement of the photographing device according to the feature vector generated by the feature vector generating unit (131), on the basis of an association between the feature vector and the movement of the photographing device, the association being obtained as a result of previously-executed machine learning of the feature vector and an actual movement of the photographing device. The constraint condition generating unit (130) determines the range corresponding to the information of the movement of the photographing device, the movement being identified by the motion classifying unit (such as 132).
摘要:
An optical alignment method which develops a pretilt angle by batch plane exposure without tilting a substrate. This optical alignment method provides liquid crystal aligning capability to the surface of a polymer film by exposing the surface of the polymer film through a slit exposure mask while the surface of the polymer film and the slit exposure mask are moved relative to each other substantially at a fixed rate. Alternatively, liquid crystal aligning capability is provided to the surface of the polymer film by exposing the surface of the polymer film through an optical exposure pattern while the optical exposure pattern having a plurality of lines with a certain width at certain intervals is formed on the surface of the polymer film continuously.
摘要:
It is intended to provide a method for removing leukocytes whereby leukocytes can be removed at a higher level of performance in a short filtration time without clogging. A method comprising passing a leukocyte-containing liquid through a filter for removing leukocytes made of a nonwoven fabric having an average fiber diameter of from 0.3 to 3.0 μm and thus removing the leukocytes from the leukocyte-containing liquid to give a leukocyte-free liquid, characterized by using a nonwoven fabric having a formation index (y) of 50 or less corresponding to a thickness of 0.3 mm; use of a leukocyte removal filter therefor; and the leukocyte removal filter.
摘要:
An optical isolator comprises an optical element assembly (10) including a Faraday rotator (1) and a pair of polarizing elements (2) one of which is arranged on a light-input side of the Faraday rotator (1) to serve as a polarizer and other of which is arranged on a light-output side to serve as an analyzer. The optical element assembly (10) is received at the center of an inner bore or a cavity of a cylindrical iron-chromium-cobalt (Fe—Cr—Co) magnet (5). The Faraday rotator (1) comprises a magnetic garnet thick film which is a Bi garnet thick film remarkably low in saturation magnetization as compared with a rare-earth garnet. The Fe—Cr—Co magnet (5) serves as a field-application magnet for generating a magnetic field applied to magnetize the Bi garnet thick film and also as an isolator casing because it is excellent in magnetic and mechanical characteristics.
摘要:
A learning pattern is classified into a plurality of patterns by a classifying means. The result of the classification, as well as the representative feature quantity calculated by the representative feature quantity operating unit, are stored in a dictionary storing unit in respective regions. Upon recognition, after feature quantities of unknown input patterns are extracted from the feature quantity extracting unit, the similarity operating unit arithmetically operates the similarity between the input pattern and respective learning patterns for respective regions. The point accumulating means searches a dictionary storing unit and accumulates points corresponding to a rank for respective learning patterns included in class having the upper rank from the highest to a lower similarity for respective regions. The recognition result candidate outputting unit outputs the recognition result candidate in accordance with the order of the higher accumulated point for the learning pattern. Therefore, the present pattern recognition can decrease the memory capacity of the dictionary and the number of the calculations.
摘要:
This invention pertains to a data processing system for pattern recognition by sorting the scores of the candidate patterns by their feature vectors per an associative matching method. It aims first at determining feature vectors at a high speed even if a recognition device obtains dictionary data in dot units, second at determining scores by an association conformance recognition device, and third at sorting inputted data from the highest score at a high speed. The feature of this invention resides in a data processing device that incorporates a first memory for memorizing stroke information for patterns supplied in dot units; a first address generator for generating first addresses for specifying the positions at which the stroke information is memorized; a second memory for memorizing weight data for the stroke information; a second address generator for generating second addresses for reading stroke information memorized in the first memory and third addresses for reading, from the second memory, the weight data corresponding to the stroke information read by the second addresses; accumulators of the number of the stroke directions supplied with weight data outputted from the second memory; and a decoder for decoding the stroke information memorized in the first memory and for accumulating weights for the respective stroke directions to enable the accumulators per the decoding result.