Abstract: An electronic device includes a graphic processor and a memory device. The graphic processor includes an artificial neural network engine that makes an object recognition model learn by using learning data and weights to provide a learned object recognition model. The memory device, divides a feature vector into a first sub feature vector and a second feature vector, and performs a first calculation to apply the second sub feature vector and the weights to the learned object recognition model to provide a second object recognition result. The artificial neural network engine performs a second calculation to apply the first sub feature vector and the weights to the learned object recognition model to provide a first object recognition result and provides the first object recognition result to the memory device. The second calculation is performed in parallel with the first calculation.

Abstract: An electronic device includes a graphic processor and a memory device. The graphic processor includes an artificial neural network engine that makes an object recognition model learn by using learning data and weights to provide a learned object recognition model. The memory device divides a feature vector into a first sub feature vector and a second feature vector, and performs a first calculation to apply the second sub feature vector and the weights to the learned object recognition model to provide a second object recognition result. The artificial neural network engine performs a second calculation to apply the first sub feature vector and the weights to the learned object recognition model to provide a first object recognition result and provides the first object recognition result to the memory device. The second calculation is performed in parallel with the first calculation.

Abstract: A vehicle may include a side mirror including a housing and a mirror member provided inside the housing and rotatable in a first axis and a second axis in the housing; an input unit configured to receive an identification information related to a driver; a driving apparatus configured to identify a target position corresponding to the identification information related to the driver, acquire a plurality of paths for moving the mirror member between the identified target position and the current position, identify each of positions for switching the movement direction of the acquired plurality of paths, select a path having positions for switching the movement direction closest to a predetermined center position among the respective positions for switching the movement direction thereof, and selectively control rotation of the first axis and rotation of the second axis of the mirror member based on the selected path.

Abstract: An electronic device includes a graphic processor and a memory device. The graphic processor includes an artificial neural network engine that makes an object recognition model learn by using learning data and weights to provide a learned object recognition model. The memory device divides a feature vector into a first sub feature vector and a second feature vector, and performs a first calculation to apply the second sub feature vector and the weights to the learned object recognition model to provide a second object recognition result. The artificial neural network engine performs a second calculation to apply the first sub feature vector and the weights to the learned object recognition model to provide a first object recognition result and provides the first object recognition result to the memory device. The second calculation is performed in parallel with the first calculation.

Abstract: An electronic device includes a graphic processor and a memory device. The graphic processor includes an artificial neural network engine that makes an object recognition model learn by using learning data and weights to provide a learned object recognition model. The memory device, divides a feature vector into a first sub feature vector and a second feature vector, and performs a first calculation to apply the second sub feature vector and the weights to the learned object recognition model to provide a second object recognition result. The artificial neural network engine performs a second calculation to apply the first sub feature vector and the weights to the learned object recognition model to provide a first object recognition result and provides the first object recognition result to the memory device. The second calculation is performed in parallel with the first calculation.

Abstract: A vehicle may include a side mirror including a housing and a mirror member provided inside the housing and rotatable in a first axis and a second axis in the housing; an input unit configured to receive an identification information related to a driver; a driving apparatus configured to identify a target position corresponding to the identification information related to the driver, acquire a plurality of paths for moving the mirror member between the identified target position and the current position, identify each of positions for switching the movement direction of the acquired plurality of paths, select a path having positions for switching the movement direction closest to a predetermined center position among the respective positions for switching the movement direction thereof, and selectively control rotation of the first axis and rotation of the second axis of the mirror member based on the selected path.

Abstract: An electronic device includes a graphic processor and a memory device. The graphic processor includes an artificial neural network engine that makes an object recognition model learn by using learning data and weights to provide a learned object recognition model. The memory device, divides a feature vector into a first sub feature vector and a second feature vector, and performs a first calculation to apply the second sub feature vector and the weights to the learned object recognition model to provide a second object recognition result. The artificial neural network engine performs a second calculation to apply the first sub feature vector and the weights to the learned object recognition model to provide a first object recognition result and provides the first object recognition result to the memory device. The second calculation is performed in parallel with the first calculation.

Abstract: A method for measuring a substrate is provided. The method comprises irradiating a measurement beam into a site box of an identifiable pattern of a substrate, detecting a center position of the irradiated measurement beam, calculating an amount of shift of the center position of the measurement beam from the center position of the site box, and correcting the center position of the measurement beam to the center position of the site box by compensating the calculated amount of shift.

Abstract: A method for measuring a substrate is provided. The method comprises irradiating a measurement beam into a site box of an identifiable pattern of a substrate, detecting a center position of the irradiated measurement beam, calculating an amount of shift of the center position of the measurement beam from the center position of the site box, and correcting the center position of the measurement beam to the center position of the site box by compensating the calculated amount of shift.

Abstract: A load port module includes a plate to support at least one substrate, a sensor to detect foreign matter at a predetermined location, and a cleaner to remove the foreign matter at the predetermined location when the foreign matter is detected by the sensor. The cleaner removes the foreign matter at the predetermined location in synchronism with movement of the plate. A mover to move the plate continues to operate after detection of the foreign matter by the sensor and removal of the foreign matter by the cleaner.

Abstract: The present invention relates to a method of extinguishing an electric arc, which occurs in low or high voltage switchgears, by pulse discharge. The electric arc is cut like a “fuse” by connecting a condenser (201) to both ends of the electric arc at the moment when it occurs and discharging the condenser (201) through the electric arc. The condenser is charged using a resistor (202) and a diode (203), and its discharge is adjusted by some auxiliary contacts.

Abstract: A load port module includes a plate to support at least one substrate, a sensor to detect foreign matter at a predetermined location, and a cleaner to remove the foreign matter at the predetermined location when the foreign matter is detected by the sensor. The cleaner removes the foreign matter at the predetermined location in synchronism with movement of the plate. A mover to move the plate continues to operate after detection of the foreign matter by the sensor and removal of the foreign matter by the cleaner.

Abstract: An apparatus and a method for removing blooming of a camera image are provided. When a lane marking is not normally detected from an image obtained by photographing a front view or a rear view of a car, it is determined whether the blooming is generated in the image. When it is determined that the blooming is generated, a mask for removing the blooming is generated and superposed on the image. Accordingly, a situation that the lane marking is not normally detected due to the blooming caused by a head light of a car can be reduced or prevented in advance.

Abstract: An apparatus and method for correcting the backlight of a camera are disclosed. When the camera for photographing a moving image by a photographing device, such as a general-purpose CCD or a general-purpose CMOS, enters into a backlight shooting state for simultaneously photographing an object with a very high luminance level and an object with a very low luminance level, as the shutter speed of the photographing device is varied to a long shutter speed and a short shutter speed, an image shot with the long shutter speed and an image shot with the short shutter speed are synthesized and outputted, thereby enabling a proper luminance level object photographing operation for each of the objects. Accordingly, it is possible to efficiently prevent the phenomenon that the object with the low luminance level looks dark due to the object with the high luminance level.

Abstract: An apparatus and method for controlling an image sensor output are discussed. When a certain number of valid pixels of an image sensor, such as a CCD, is selected, a corresponding timing parameter is calculated. Signal values output from charge wells respectively corresponding to unit cells of the image sensor are separated or merged and then output by using the calculated timing parameter. Therefore, an image of a user-defined number of valid pixels can be varied and output in various ways, and all electric charges that are photoelectrically converted by the maximum number of valid pixels of the image sensor can be efficiently used.

Abstract: An apparatus and a method for removing blooming of a camera image are provided. When a lane marking is not normally detected from an image obtained by photographing a front view or a rear view of a car, it is determined whether the blooming is generated in the image. When it is determined that the blooming is generated, a mask for removing the blooming is generated and superposed on the image. Accordingly, a situation that the lane marking is not normally detected due to the blooming caused by a head light of a car can be reduced or prevented in advance.

Abstract: An apparatus and method for controlling an image sensor output are discussed. When a certain number of valid pixels of an image sensor, such as a CCD, is selected, a corresponding timing parameter is calculated. Signal values output from charge wells respectively corresponding to unit cells of the image sensor are separated or merged and then output by using the calculated timing parameter. Therefore, an image of a user-defined number of valid pixels can be varied and output in various ways, and all electric charges that are photoelectrically converted by the maximum number of valid pixels of the image sensor can be efficiently used.

Abstract: A method and an apparatus for processing images in a camera are provided. Even lines and odd lines of a photographing device are respectively read at different points of time to obtain even line data and odd line data and motion estimation is performed on one of the even line data and the odd line data on the basis of an image related to the other of the even line data and the odd line data right of an immediately prior point in time to the one of the even line data and the odd line data to generate a motion-compensated image. Accordingly, a quantity of electric charges charged in the photographing device comes to be increased such that a bright image of an object can be taken and a motion of an image comes to be accurately estimated such that a more distinct moving image can be obtained.

Abstract: An apparatus and method for correcting the backlight of a camera are disclosed. When the camera for photographing a moving image by a photographing device, such as a general-purpose CCD or a general-purpose CMOS, enters into a backlight shooting state for simultaneously photographing an object with a very high luminance level and an object with a very low luminance level, as the shutter speed of the photographing device is varied to a long shutter speed and a short shutter speed, an image shot with the long shutter speed and an image shot with the short shutter speed are synthesized and outputted, thereby enabling a proper luminance level object photographing operation for each of the objects. Accordingly, it is possible to efficiently prevent the phenomenon that the object with the low luminance level looks dark due to the object with the high luminance level.