Abstract: The present invention, according to one embodiment, in order to generate a descriptor for a feature point extracted from an input image, may: determine the size of each of a plurality of sub-patches included in a main patch having a preset size on the basis of the feature point in the input image; calculate gradient directions for each pixel constituting a target sub-patch from among the plurality of sub-patches; determine the number of bins for distinguishing the gradient directions; generate, with respect to the target sub-patch, a histogram for a plurality of bins on the basis of the gradient directions of the pixels constituting the target sub-patch; and generate a descriptor for the feature point on the basis of the histogram. Various other embodiments may also be possible.

Abstract: A method and an apparatus for training an image-enhanced neural network model are provided. According to one embodiment, the training method may comprise the steps of: acquiring sample images having various image qualities; generating enhanced images of at least some of the sample images by using image enhancement software having an image enhancement function; constructing, from the sample images and the enhanced images, training data that forms pairs of input data and target data; using the training data so as to output an enhanced output image in response to input of a low-image-quality input image; and performing supervised training on a first neural network model for outputting a corresponding-image-quality output image in response to input of a high-image-quality input image.

Abstract: The present invention may comprise: in order to receive an analog image signal, generating a camera initialization signal including operation period information of a camera; transmitting the camera initialization signal to a plurality of analog image signal transmission apparatuses; determining whether or not each of the plurality of analog image signal transmission apparatuses is ready to transmit the analog image signal; if determining that each of the plurality of analog image signal transmission apparatuses is ready to transmit the analog image signal, simultaneously transmitting a transmission command signal to the plurality of analog image signal transmission apparatuses; and receiving a plurality of analog image signals from the plurality of analog image signal transmission apparatuses.

Abstract: Provided is an object detection method including detecting a first object by using an object detection filter in an original image and images of which resolutions are downscaled and detecting a second object by using a neural network-based object detection model of which an input is a target image that is determined to be one of the original image and the images of which resolutions are downscaled.

Abstract: In order to process an image according an embodiment: first detailed feature information may be generated by processing an original image; second detailed feature information may be generated by processing the first detailed feature information; lane type context information may be generated by processing the original image; a type of a target lane in the original image may be determined on the basis of the second detailed feature information and the lane type context information; lane instance context information may be generated by processing the first detailed feature information; and an instance of the target lane may be determined on the basis of the second detailed feature information and the lane instance context information.

Abstract: In order to determine whether bumps and depressions are present on the ground in a target image, the present invention: acquires a V-disparity map on the basis of a target image acquired by means of an image sensor; detects, in the V-disparity map, at least one component indicating information about the ground in the target image; determines a polynomial in a coordinate system in the V-disparity map for the component; determines a straight line passing through two different inflection points on a curve, which corresponds to the polynomial, in the coordinate system in the V-disparity map; determines a target point on the curve, the target point having a tangential slope identical to the slope of the straight line; calculates the distance between the straight line and the target point; and determines, on the basis of the distance, whether the ground includes bumps and depressions.

Abstract: A distance determination method includes generating a target image using a camera, generating a lane plane image based on the target image, generating a depth map based on the target image, and determining a target distance between the camera and a point in a real world corresponding to a target pixel in the lane plane image based on the lane plane image and the depth map.

Abstract: A method of transmitting data, performed by a transmission device of a serial isolated communication system includes obtaining data to be transmitted, generating a plurality of data sets having 2 bits based on the data, determining a plurality of data pulse signals by encoding the plurality of data sets according to a preset rule, and transmitting the plurality of data pulse signals to a reception device of the serial isolated communication system. Each of the plurality of data pulse signals is composed of any two pulses of a first positive pulse having a first amplitude, a second positive pulse having a second amplitude, a first negative pulse having the first amplitude, or a second negative pulse having the second amplitude.

Abstract: Provided is a color filter array. According to one embodiment, the color filter array may comprise a repeating series of 2*2 cell structures including one red pixel, two green pixels, and one clear pixel, and the one red pixel and the one clear pixel in the 2*2 cell structure may be arranged in a first diagonal direction and the two green pixels may be arranged in a second diagonal direction.

Abstract: To detect an object, an embodiment may: calculate a first distance to a first region and a second distance to a second region of a target object in a first image captured from a first point of view; calculate a third distance to the first region and a fourth distance to the second region of the target object in a second image captured from a second point of view; calculate a first difference between the first distance and the third distance; calculate a second difference between the second distance and the fourth distance; and determine whether the target object is a three-dimensional object, on the basis of the first difference and the second difference.

Abstract: The present invention, according to one embodiment, in order to generate a descriptor for a feature point extracted from an input image, may: determine the size of each of a plurality of sub-patches included in a main patch having a preset size on the basis of the feature point in the input image; calculate gradient directions for each pixel constituting a target sub-patch from among the plurality of sub-patches; determine the number of bins for distinguishing the gradient directions; generate, with respect to the target sub-patch, a histogram for a plurality of bins on the basis of the gradient directions of the pixels constituting the target sub-patch; and generate a descriptor for the feature point on the basis of the histogram. Various other embodiments may also be possible.

Abstract: In order to determine whether a reset function of an electronic device is safe, the states of a first checker and a second checker are toggled to preset states when a reset root is received from an external device. When a release of a local reset is received, whether the local reset was normally released is determined on the basis of a system clock. When the local reset is not normally released, the state of the second checker is toggled, and when the final local reset among the local resets is normally released, the state of the first checker is toggled. The release states of the local resets are monitored on the basis of the state of the first checker and the state of the second checker.

Abstract: In order to determine whether a reset function of an electronic device is safe, the states of a first checker and a second checker are toggled to preset states when a reset root is received from an external device. When a release of a local reset is received, whether the local reset was normally released is determined on the basis of a system clock. When the local reset is not normally released, the state of the second checker is toggled, and when the final local reset among the local resets is normally released, the state of the first checker is toggled. The release states of the local resets are monitored on the basis of the state of the first checker and the state of the second checker.

Abstract: In order to determine whether bumps and depressions are present on the ground in a target image, the present invention: acquires a V-disparity map on the basis of a target image acquired by means of an image sensor; detects, in the V-disparity map, at least one component indicating information about the ground in the target image; determines a polynomial in a coordinate system in the V-disparity map for the component; determines a straight line passing through two different inflection points on a curve, which corresponds to the polynomial, in the coordinate system in the V-disparity map; determines a target point on the curve, the target point having a tangential slope identical to the slope of the straight line; calculates the distance between the straight line and the target point; and determines, on the basis of the distance, whether the ground includes bumps and depressions.

Abstract: Provided is an object detection method including detecting a first object by using an object detection filter in an original image and images of which resolutions are downscaled and detecting a second object by using a neural network-based object detection model of which an input is a target image that is determined to be one of the original image and the images of which resolutions are downscaled.

Abstract: A method and apparatus for exchanging a protocol is provided. A protocol may be indicated in a preset section of a vertical blanking interval (VBI) of an analog video and thereby transmitted. A protocol transmission apparatus may transmit an analog video signal in which a protocol signal is indicated to a protocol reception apparatus. The protocol reception apparatus may detect a protocol by receiving the analog video signal from the protocol transmission apparatus.

Abstract: Provided is a method and apparatus for converting, to a digital video signal, an analog video signal capable of having a variety of resolutions. A video reception apparatus determines a format of an analog video signal by sampling the analog video signal. The video reception apparatus converts the analog video signal to a digital video signal based on the determined format. A comb filter is used to separate a color difference signal and a luminance signal from the digital video signal, and the digital video signal is reconfigured using the luminance signal and the color difference signal.

Abstract: Provided is a method and apparatus for transmitting a video signal. The video signal transmission method and apparatus receives a digital video signal from a video sensor, reconfigures the digital video signal by adaptively adjusting at least one of the number of horizontal blank samples and the number of vertical blank lines of the digital video signal so that the digital video signal corresponds to a video standard, converts the reconfigured digital video signal to an analog video signal, and transmits the analog video signal.

Abstract: The present invention relates to an apparatus and method for recovering images damaged by weather phenomena, including: an input unit for receiving an image; a damaged image detection unit for detecting an image damaged by weather phenomena; an image recovery unit for recovering the damaged image; and an output unit for outputting the recovered image, wherein the damaged image recovery unit can recover the contrast and/or color of the damaged image.

Abstract: With regard to a distortion correction apparatus and method which are applicable to hardware for real-time distortion correction, provided are a distorted image correction apparatus and method for correcting a distortion of an image by using a non-polynomial estimation function suitable for a hardware operation, and for correcting the distortion of the image by utilizing a morphing form.