Abstract: Provided is an image processing apparatus including a memory storing one or more instructions and a processor configured to execute the one or more instructions stored in the memory, wherein the processor is further configured to execute the one or more instructions to generate a second image by performing a deconvolution operation on a first image and a kernel comprising one or more weights, set values of the one or more weights based on the second image, and adjust the values of the one or more weights based on positions of the one or more weights in the kernel.

Abstract: A three-dimensional data encoding method includes: (i) in a first case where a layered structure is generated by classifying three-dimensional points into layers: encoding attribute information for the three-dimensional points based on the layered structure; and generating a bitstream including layer information utilized for the generation of the layered structure; and (ii) in a second case where the three-dimensional points are not classified: encoding attribute information for the three-dimensional points; and generating a bitstream not including the layer information.

Abstract: A system and method of calibrating moving cameras capturing a sporting event is disclosed herein. A computing system retrieves a broadcast video feed for a sporting event. The broadcast video feed includes a plurality of video frames. The computing system labels, via a neural network, components of a playing surface captured in each video frame. The computing system matches a subset of labeled video frames to a set of templates with various camera perspectives. The computing system fits a playing surface model to the set of labeled video frames that were matched to the set of templates. The computing system identifies camera motion in each video frame using an optical flow model. The computing system generates a homography matrix for each video frame based on the fitted playing surface model and camera motion. The computing system calibrates each camera based on the homography matrix generated for each video frame.

Abstract: An online fitting method and apparatus receive, from a user, a body size of the user and a target size of clothes desired by the user for fitting, obtain barycentric coordinate information corresponding to a result of fitting the clothes of the target size to a reference avatar selected based on the body size of the user, generate a target avatar having the same mesh topology as the reference avatar and corresponding to the body size of the user, fit the clothes to the target avatar by applying the barycentric coordinate information to the target avatar, and display a result of fitting the clothes to the target avatar.

Abstract: The present invention acquires a local enlargement ratio in an image transformed based on a transformation parameter and then acquires a transformed image which is a captured image transformed based on the transformation parameter. The present invention acquires a sharpening parameter for sharpening the transformed image based on the lens information, which indicates the resolution of a lens used at the time of capturing the captured image, and the enlargement ratio, and performs sharpening of the transformed image based on the sharpening parameter.

Abstract: Disclosed is an image sensing device and an operating method thereof. The image sensing device may include an analyzer suitable for analyzing a state of each of multiple kernels based on system information and a plurality of pixel values, a detector suitable for detecting color noise of a target pixel value among pixel values included in a target kernel among the multiple kernels, according to the analysis result of the analyzer, and a corrector suitable for correcting the target pixel value according to the detection result of the detector.

Abstract: A process for linking related images and videos is disclosed. The process can include receiving listing information including the images and the video, processing the images using an image processor to determine one or more image descriptors, processing the video using a video processor to determine video descriptors, comparing the image descriptors to the video descriptors, calculating a similarity value for each of the image descriptors in relation to each of the video descriptors, and linking the videos and the images based upon the calculated similarity value. The method can also include causing the display of a user interface including a video linking user interface element operable to cause playback of a relevant portion of the video linked with the displayed image and an image linking user interface element operable to cause display of a relevant image linked with a portion of the video playback.

Abstract: A camera device according to an embodiment may include: an image sensor which generates first Bayer data having a first resolution; and a processor which performs deep learning on the basis of the first Bayer data to output second Bayer data having a second resolution higher than the first resolution.

Abstract: An unmanned mobile observation device, such as an unmanned aerial vehicle, can be used to acquire observation data for a sensitive area. The observation data can include image data and geolocation data corresponding to a real-world geographic location at which the image data was acquired. The geolocation data can be accurate to five centimeters or less. The observation data can be evaluated to identify any defects within the sensitive area. Such evaluation can include comparison with previously acquired image data for the sensitive area, which can be enables by using the geolocation data to select the corresponding previously acquired image data.

Abstract: A method of processing a first image in a first plurality of images, wherein the first image is divided into a plurality of pixel blocks, is proposed, which comprises, for a current block of the first image: selecting, in a set of a plurality of predefined interpolation filters, an interpolation filter based on a prediction of an interpolation filter determined by a supervised learning algorithm to which data related to the current block is input; and using the selected interpolation filter for calculating fractional pixel values in a second image of the plurality of images for a temporal prediction of pixels of the current block based on a reference block correlated to the current block in the second image, wherein the second image is distinct from the first image and was previously encoded according to an image encoding sequence for encoding the images of the plurality of images.

Abstract: Systems and methods are provided for determining whether an image or location contains a barcode. The systems and methods include comparing visual data extracted from an image to target visual data. The systems and methods also include comparing different portions of captured barcode data to corresponding portions of target barcode data. The systems and methods also include utilizing barcode formatting data when comparing or computing a similarity measure. The systems and methods also include causing a second image to be captured when no match is found in a first image. The systems and methods also include identifying a target barcode from a barcode database based on spatial information about where an image was captured. The systems and methods also include validating a barcode based on different regions of the barcode from two different images of the barcode. The systems and methods also include using two barcodes captured in an image.

Abstract: This application relates to a system for compressing captured images. The system may include an image capturing device configured to continuously capture images, for example, of a location, or frame, or field of view, and convert the captured images respectively into a plurality of sets of digital data, each of the plurality of sets of digital data including at least two consecutive rows of digital data. The system may also include a processor circuit configured to combine the at least two sets of the plurality of sets of digital data to generate a pre-compression set of digital data and simultaneously compress the pre-compression set of digital data simultaneously or together in a single compression operation.

Abstract: An active sensor for performing active measurements of a scene is presented. The active sensor includes at least one transmitter configured to emit light pulses toward at least one target object in the scene, wherein the at least one target object is recognized in an image acquired by a passive sensor; at least one receiver configured to detect light pulses reflected from the at least one target object; a controller configured to control an energy level, a direction, and a timing of each light pulse emitted by the transmitter, wherein the controller is further configured to control at least the direction for detecting each of the reflected light pulses; and a distance measurement circuit configured to measure a distance to each of the at least one target object based on the emitted light pulses and the detected light pulses.

Abstract: Disclosed is an image sensing device including a first module suitable for generating a plurality of interpolated images separated for each color channel, based on a raw image and a plurality of first convolution layers, a second module suitable for generating a plurality of refined images separated for each color channel, based on the plurality of interpolated images and a plurality of second convolution layers, and a third module suitable for generating at least one output image corresponding to the raw image, based on the plurality of refined images and a plurality of third convolution layers.

Abstract: A system and method for determining a correspondence map between a first and second image by determining a set of correspondence vectors for each pixel in the first image and selecting a correspondence vector from the set of correspondence vectors based on a cost value.

Abstract: Multiple snapshots of a scene are captured within an executing application (e.g., a video game). When each snapshot is captured, associated color values per pixel and a distance or depth value z per pixel are stored. The depth information from the snapshots is accessed, and a point cloud representing the depth information is constructed. A mesh structure is constructed from the point cloud. The light field(s) on the surface(s) of the mesh structure are calculated. A surface light field is represented as a texture. A renderer uses the surface light field with geometry information to reproduce the scene captured in the snapshots. The reproduced scene can be manipulated and viewed from different perspectives.

Abstract: Disclosed is a system and associated methods for generating a point-accurate three-dimensional (“3D”) model of a scene with point-accurate color information from a non-cosited capture of the scene. The system may define a first model based on positional measurements obtained from a first device. The system may capture images of the scene with a different set of devices, and may generate a second model from a photogrammetry modeling of the images. The system may align the models in order to determine positioning of the set of devices relative to the first model, and may select a particular device that captures descriptive characteristics from a position in the scene that is represented by the positional elements of a particular data point. The system may define non-positional elements of the particular data point based on the descriptive characteristics captured by the particular device.

Abstract: The present disclosure discloses a method and a device for converting two-dimensional (2D) images into three-dimensional (3D) images and a 3D imaging system, wherein the method comprises the following steps: acquiring 2D image to be processed; performing perspective transformation on the 2D image to obtain a left-eye image and a right-eye image respectively; adjusting a distance between the left-eye image and the right-eye image according to the result of perspective transformation; and synthesizing the left-eye image and the right-eye image after the distance adjustment. In embodiments of the present disclosure, binocular parallax images are created by performing perspective transformation on the 2D image to be processed; the distance between the left-eye image and the right-eye image after perspective transformation is adjusted to form binocular parallax and create a convergence angle, so that the images observed by naked eyes are located at different depths, thus different stereoscopic effects may be seen.

Abstract: An apparatus to facilitate deep learning based selection of samples for adaptive supersampling is disclosed. The apparatus includes one or more processing elements to: receive training data comprising input tiles and corresponding supersampling values for the input tiles, wherein each input tile comprises a plurality of pixels, and train, based on the training data, a machine learning model to identify a level of supersampling for a rendered tile of pixels.

Abstract: An apparatus for decoding an encoded audio signal to obtain a reconstructed audio signal includes a receiving interface for receiving one or more frames comprising information on a plurality of audio signal samples of an audio signal spectrum of the encoded audio signal, and a processor for generating the reconstructed audio signal. The processor is configured to generate the reconstructed audio signal by fading a modified spectrum to a target spectrum, if a current frame is not received by the receiving interface or if the current frame is received by the receiving interface but is corrupted, wherein the modified spectrum includes a plurality of modified signal samples, wherein, for each of the modified signal samples of the modified spectrum, an absolute value of the modified signal sample is equal to an absolute value of one of the audio signal samples of the audio signal spectrum.

Abstract: High dynamic range (HDR) support is provided for legacy application programs, such as games that are configured to display standard dynamic range (SDR) frames. HDR frames may be generated without modifying the legacy application program. The buffer creation process of the legacy application program is intercepted and modified before creation of the SDR format buffer so that the buffer is configured to use an upgraded SDR format having an increased bit depth compared with a conventional SDR buffer. Rather than tone mapping and quantizing rendered image data to the lower bit depth for storage in the conventional SDR buffer, the rendered image data is tone mapped and quantized for storage at the increased bit depth of the upgraded SDR buffer. Therefore, the luminance and greater dynamic range of the tone mapped data is better preserved compared with outputting conventional SDR frames.

Abstract: An embodiment of the present disclosure discloses an intra-prediction-based encoding and decoding method, apparatus and filter, and belongs to a video coding technical field.

Abstract: A method for interpolating values of an attribute for an image grid may include determining a root value of the attribute for a root node located centrally in the image grid, pre-calculating metadata for multiple child nodes in one or more hierarchical levels based on one or more gradients of the attribute, and deriving values of the attribute for each of the child nodes at each of the hierarchical levels based on the corresponding root value and metadata for the hierarchical level of each child node, wherein each child node may be used as a root node in the next hierarchical level. The image grid may have multiple outer cells arranged radially around a central cell, and the root node may be located in the central cell.

Abstract: A method is described to greatly improve the efficiency of and reduce the complexity of image compression when using single-sensor color imagers for video acquisition. The method in addition allows for this new image compression type to be compatible with existing video processing tools, improving the workflow for film and television production.

Abstract: A chromaticity compensation method, a chromaticity compensation device and a storage medium are disclosed. The chromaticity compensation method includes that, according to a preset standard chromaticity value and measured chromaticity values corresponding to a plurality of sub-display regions in a display region, chromaticity compensation values of the sub-display regions are determined, and the display region is divided into the plurality of sub-display regions in close arrangement, and each of the sub-display regions includes a plurality of sub-pixels; a data smoothing treatment is performed between the chromaticity compensation values of the sub-display regions that are adjacent, so as to determine the smoothed chromaticity compensation values of the sub-pixels in the sub-display regions; and upon display, chromaticity compensation of the sub-pixels are performed by adopting the smoothed chromaticity compensation values.

Abstract: A processor generates a depth map from two images, including a high-dynamic range (HDR) image. The processor receives, from a first image sensor, one or more first images, wherein the one or more first images have a first field-of-view (FOV) and a first resolution, and receives, from a second image sensor, one or more second images, wherein the one or more second images have a second FOV and a second resolution. The processor generates a first HDR image from the one or more first images, and performs tone alignment on the one or more second images based on the first HDR image to produce a second HDR image. The processor generates a depth map using the first HDR image and the second HDR image. The processor may use the depth map to apply a bokeh effect to the first HDR image.

Abstract: An encoder includes processing circuitry and memory. Using the memory, the processing circuitry: encodes and reconstructs an image to generate a reconstructed image; determines, according to a characteristic of a block in the reconstructed image, an interpolation method for interpolating pixels located outside a referable region including the block; interpolates the pixels located outside the referable region, using the interpolation method determined; and applies a filter to the block using the pixels interpolated.

Abstract: A memory-efficient filtering approach is used to code images and video. A buffer having a fixed size based on a size of processing units to use for filtering a video frame is allocated. For each of the processing units, pre-filtered pixel values are copied from a respective region of the video frame to the buffer based on a writing point for the video frame and an offset applied to the writing point, filtering is performed against the pre-filtered pixel values from the buffer to produce filtered pixel values, and the filtered pixel values are written to the video frame based on the writing point and the offset. The filtering may be performed using a loop restoration tool, such as where the pre-filtered pixel values are output from a constrained directional enhancement filter (CDEF) tool. Alternatively, the filtering may be performed using the CDEF tool or another coding tool.

Abstract: An inspection system is disclosed. In one embodiment, the inspection system includes an interferometer sub-system configured to acquire an interferogram of a sample. The inspection system may further include a controller communicatively coupled to the interferometer sub-system. The controller is configured to: receive the interferogram from the interferometer sub-system; generate a phase map of the sample based on the received interferogram, wherein the phase map includes a plurality of pixels; select a sub-set of pixels of the plurality of pixels of the phase map to be used for phase unwrapping procedures; perform one or more phase unwrapping procedures on the sub-set of pixels of the phase map to generate an unwrapped phase map; and generate a surface height map of the sample based on the unwrapped phase map.

Abstract: A video processing method is provided. In the method, circuitry of a terminal determines a second video frame portion associated with a second time point that is after a first time point associated with a first video frame portion of a video. The circuitry generates a transitional video frame based on the first video frame portion and the second video frame portion. A color value of a pixel at a target pixel location in the transitional video frame is within a target color interval. The target color interval is determined according to a color value of a pixel at the target pixel location in the first video frame portion and a color value of a pixel at the target pixel location in the second video frame portion. The circuitry performs display control of the video according to the transitional video frame.

Abstract: A Bayer-mask image is decoded by forming a decoded green array; calculating a slope at each pixel of the array, expressed as an angle; calculating an activity at each pixel; converting the slope angle and the activity into a complex number for each pixel, of modulus equal to the activity and argument equal to twice the slope angle; expressing said complex numbers in Cartesian coordinates to form a Cartesian slope signal and filtering the Cartesian slope signal with a linear spatial filter to derive a slope measure. Blue-green and red-green values are calculated and interpolated using a slope-adaptive interpolation filter steered by said slope measure.

Abstract: Disclosed is a display apparatus including: a signal receiver configured to receive an image signal; a display configured to display an image; a processor configured to: calculate a change degree and a change direction of pixel value differences between at least one first pixel and two or more second pixels of an image, and change a pixel value of the first pixel based on the pixel value difference which is relatively small among the pixel value differences obtained by the calculated change degree and the calculated change direction. According to this, it is possible to enhance the image details without generating or and increasing of the noises.

Abstract: A method and apparatus for video encoding or decoding performed by a video encoder or a video decoder incorporating advanced multiple transform (AMT) are disclosed. According to this method, a transform set consisting of two type candidates designated as a first type and a second type is determined, where the first type and the second type are different. If AMT is used by the current block, a vertical transform is selected from the transform set and a horizontal transform is selected from the transform set. A transform index, for indicating both vertical transform selection and horizontal transform selection, can be signaled at the encoder side or the transform index is parsed at the decoder side. Another method and apparatus incorporating AMT are also disclosed where the transform set is determined according to the resolution of current MVD.

Abstract: A Bayer-mask is decoded by forming a decoded green array, calculating red green and blue-green color difference signals, deriving a first full-resolution grid of each colour difference signal using a first interpolation filter; deriving a second full-resolution grid of each colour difference signal using a second interpolation filter taking a larger number of colour difference signal inputs than the first; generating a cross-colour weighting signal; and forming a mixed colour difference signal from the first and second full-resolution grids in accordance with the cross-colour weighting signal.

Abstract: Systems and methods are provided for image scaling with efficient pixel resolution conversion. One embodiment is an image scaling system that includes a processor that receives an instruction to scale an input image having a first resolution by a scaling ratio to create an output image having a second resolution, and determines a rational fraction approximation of the scaling ratio to obtain a numerator integer representing a first block of consecutive pixels in the output image and a denominator integer representing a second block of consecutive pixels in the input image. The processor also calculates pixel values of the first block of pixels by: bit shifting pixel values of the second block of pixels, obtaining a sum of bit shifting operations of neighboring pixel values of the second block of pixels, and correcting the sum with a correction value to compensate for rounding errors in the bit shifting.

Abstract: In one embodiment, an electronic device is provided, comprising: at least one non-transitory memory; a touch screen; a camera; and one or more processors in communication with the at least one non-transitory memory, the touch screen, and the camera, wherein the one or more processors execute instructions stored in the non-transitory memory to cause the apparatus to: display, via the touch screen, a first virtual display layer including contents; display, via the touch screen, at least one user interface element; detect, via the touch screen, at least a portion of touch on the least one user interface element; when an aspect of the touch is detected to surpass a threshold, display, via the touch screen, a plurality of markings in a second virtual display layer that appears to have a lesser depth than the first virtual display layer, where at least a portion of the second virtual display layer is at least partially translucent so that at least a portion of the contents of the first virtual display layer is visi

Abstract: An image processing method includes determining whether image data acquired by reading a document with a reading unit is color or monochrome, saving color image data determined to be color and monochrome image data generated by monochrome conversion of the color image data, in association with each other as a learning sample, deciding a color conversion parameter to be used when image data read by the reading unit and determined to be monochrome image data is converted into a color image, using the saved learning sample, and converting the image data read by the reading unit and determined to be the monochrome image data into color image data, using the decided color conversion parameter.

Abstract: An image processing device includes: a high-resolution combining unit that generates a high-resolution combined image by combining a standard image with at least one reference image other than the standard image in a high-resolution image space having higher resolution than the standard image or the at least one reference image, the standard image and the at least one reference image being acquired by photographing a subject in a time-series manner by using an imaging element having a plurality of types of color filters arranged for each pixel; a moving-object detecting unit that determines at least one correlation amount by calculating a magnitude relationship between a plurality of correlation values within an arbitrary region of the high-resolution combined image generated by the high-resolution combining unit; and an image correcting unit that corrects the high-resolution combined image based on the correlation amount determined by the moving-object detecting unit.

Abstract: A wide-area image acquiring method which includes capturing a global image in a preset wide-area visual area, capturing at least two partial images whose definition is preset and that are in the preset wide-area visual area, where the global image covers at least an overlap portion of view ranges of the partial images in the preset wide-area visual area, and a sum of the view ranges of the at least two partial images is greater than or equal to a view range of the global image, determining, in the global image, positions of edges of the partial images, based on same shot objects in the partial images and the global image, and performing splice processing on the at least two partial images according to the positions of the edges, to obtain a composite wide-area image of the wide-area visual area.

Abstract: A method, computer readable medium, and system are disclosed for performing spatiotemporal filtering. The method includes the steps of applying, utilizing a processor, a temporal filter of a filtering pipeline to a current image frame, using a temporal reprojection, to obtain a color and auxiliary information for each pixel within the current image frame, providing the auxiliary information for each pixel within the current image frame to one or more subsequent filters of the filtering pipeline, and creating a reconstructed image for the current image frame, utilizing the one or more subsequent filters of the filtering pipeline.

Abstract: A method comprising: receiving a source image and a target image, each depicting a dense recurring pattern comprising an array of objects arranged in close proximity to one another; applying a trained machine learning classifier to obtain a classification of each pixel in said source image and said target image into one of at least two classes; determining a pixel-level transformation between said classified source and target images, based, at least in part, on a set of transformation parameters; training a neural network to optimize said set of transformation parameters, based, at least in part, on minimizing a loss function which calculates a weighted sum of per-pixel matches between said classified source and target images; and applying said optimized set of transformation parameters to said target image. to align said target image with said source image.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for frequency based audio analysis using neural networks. One of the methods includes training a neural network that includes a plurality of neural network layers on training data, wherein the neural network is configured to receive frequency domain features of an audio sample and to process the frequency domain features to generate a neural network output for the audio sample, wherein the neural network comprises (i) a convolutional layer that is configured to map frequency domain features to logarithmic scaled frequency domain features, wherein the convolutional layer comprises one or more convolutional layer filters, and (ii) one or more other neural network layers having respective layer parameters that are configured to process the logarithmic scaled frequency domain features to generate the neural network output.

Abstract: The present invention relates to a method for performing adaptive filtering according to a block boundary, comprising the steps of: determining whether a current pixel is within a block boundary; determining the shape of a filter depending on the location of the current pixel, when the current pixel is within the block boundary; calculating a filter coefficient based on the shape of the filter; and performing an adaptive loop filtering using the filter coefficient.

Abstract: Disclosed is a method for predicting depth map coding distortion of a two-dimensional free viewpoint video, including: inputting sequences of texture maps and depth maps of two or more viewpoint stereoscopic videos; synthesizing a texture map of a first intermediate viewpoint of a current to-be-coded viewpoint and a first adjacent viewpoint, and synthesizing a texture map of a second intermediate viewpoint of the current to-be-coded viewpoint and a second adjacent viewpoint by using a view synthesis algorithm; recording a synthetic characteristic of each pixel according to the texture map and generating a distortion prediction weight; and calculating to obtain total distortion according to the synthetic characteristic and the distortion prediction weight.

Abstract: A spectral measurement method of the present invention includes: a measuring step of measuring optical spectra of some sections specified among a plurality of sections on a specimen; a scalar-value calculating step of calculating, for individual measured sections, scalar values that represent information contained in the obtained optical spectra; an interpolating step of interpolating scalar values of unmeasured sections by using the calculated scalar values and by using two types of interpolation methods; an identifying step of identifying sections in which absolute values of differences between the two scalar values interpolated by using the two types of interpolation methods are equal to or greater than a predetermined threshold; and a repeating step of re-executing steps from the measuring step to the identifying step after specifying the identified sections.

Abstract: It is determined whether an image before pixel mixture has a high frequency. In a partial image obtained by an imaging device having a Bayer array, a Gr pixel which indicates a green component and is arranged in the same row as an R pixel indicating a red component is distinguished from a Gb pixel which indicates a green component and is arranged in the same row as a B pixel indicating a blue component. The Gr pixels and the Gb pixels which are distinguished from each other are separately mixed. When an image before pixel mixture has a high-frequency component, the level of the Gr pixel and the level of the Gb pixel are different from each other after pixel mixture. It is determined whether the image before pixel mixture has a high-frequency component based on levels of the Gr pixel and Gb pixel after pixel mixture.

Abstract: Tone mapping is performed by digital image processing circuitry on a macro-pixel basis. A luminance value of a macro-pixel of a digital image in a color space is determined. The macro-pixel includes a plurality of individual pixels. Respective tone-mapping gain values of each pixel of the macro-pixel are determined based on the determined luminance value of the macro-pixel. The determined tone-mapping gains are applied to the respective pixels of the macro-pixel. The color space may be a CFA color space, such as a Bayer color space.

Abstract: An imaging apparatus may include an image sensor that includes a Bayer color filter array (CFA) of 2-by-2 cell groups each of which includes a red pixel, a blue pixel, and two green pixels, and an image signal processor for processing raw data from the Bayer CFA. The image signal processor corrects the raw data to reduce artifacts such as image contrast degradation caused by using green pixels with wide transmission spectrum.

Abstract: The present technique relates to an image processing device and an image processing method for realizing accurate depth image warping operations at the receiving side when depth images with reduced resolutions are transmitted. A parallax image generating unit reduces the resolutions of parallax images B1? and C1?. A boundary information generating unit generates boundary information about the parallax images B1? and C1?. A multiplexing unit transmits a multiplexed stream containing the parallax images B1? and C1? with the reduced resolutions and the boundary information. The present technique can be applied to encoding devices that encode glasses-free 3D images, for example.

Abstract: First and second images are captured at first and second focal lengths, the second focal length being longer than the first focal length. Element sets are defined with a first element of the first image and a corresponding second element of the second image. Element sets are identified as background if the second element thereof is more in-focus than or as in-focus as the first element. Background elements are subtracted from further analysis. Comparisons are based on relative focus, e.g. whether image elements are more or less in-focus. Measurement of absolute focus is not necessary, nor is measurement of absolute focus change; images need not be in-focus. More than two images, multiple element sets, and/or multiple categories and relative focus relationships also may be used.