Abstract: A method includes training a first model to measure the banding artefacts, training a second model to deband the image, and generating a debanded image for the image using the second model. Training the first model can include selecting a first set of first training images, generating a banding edge map for a first training image, where the map includes weights that emphasize banding edges and de-emphasize true edges in the first training image, and using the map and a luminance plane of the first training image as input to the first model. Training the second model can include selecting a second set of second training images, generating a debanded training image for a second training image, generating a banding score for the debanded training image using the first model, and using the banding score in a loss function used in training the second model.

Abstract: Methods to solve the problem of performing fusion of images acquired with two cameras with different type sensors, for example a visible (VIS) digital camera and an short wave infrared (SWIR) camera, include performing image space equalization on images acquired with the different type sensors before performing rectification and registration of such images in a fusion process.

Abstract: A scalable hardware accelerator configured to compute video quality metrics is disclosed. In some embodiments, an accelerator for video quality metrics comprises an application-specific integrated circuit that includes a buffer memory configured to store at least a portion of a reference frame of a video and at least a corresponding portion of a distorted frame of a transcoded version of the video and that includes a processing unit configured to receive data from the buffer memory and compute a perception-based video quality metric for the distorted frame with respect to the reference frame.

Abstract: Sharpness at each focus position is calculated as an individual sharpness. Thereafter, one or M (where M is a natural number equal to or greater than 2 and smaller than N) corrective sharpness are calculated and image reference values are determined from the individual sharpness. An luminance value is calculated based on the image reference values and the corrective sharpness for each pixel. An all-in-focus image is generated by combining those luminance values.

Abstract: A method for detecting a finger is provided. The method includes obtaining a plurality of digital images including a first digital image captured by a camera from a field of view containing a background and a hand including at least one finger, and obtaining an identification of pixels of the plurality of digital images that correspond to at least one finger that is visible in the plurality of digital images rather than to the background, the pixels being identified by: obtaining, from the digital images, a Gaussian brightness falloff pattern indicative of at least one finger, identifying an axis of the at least one finger based on the obtained Gaussian brightness falloff pattern indicative of the at least one finger without identifying edges of the at least one finger, and identifying the pixels that correspond to the at least one finger based on the identified axis.

Abstract: A method and apparatus for image processing are provided. A mask is obtained by separating an original image into a background image and a foreground image. A partial stylized image is obtained by transforming the background image or the foreground image according to a selected style. A stylized image is obtained according to the mask and the partial stylized image.

Abstract: Provided are a cell image evaluation device, method, and program which are capable of more accurate and high reliable evaluation even though a captured image of each part to be observed within a container deteriorates. The cell image evaluation device includes an image evaluation unit that evaluates a state of a cell included in a captured image obtained by capturing an inside of the container that contains the cell based on the captured image, and a deterioration determination unit that determines whether or not the captured image deteriorates. The image evaluation unit changes an evaluation method of the captured image according to a determination result of the deterioration determination unit.

Abstract: A method of digitally processing an image comprises: generating an intensity distribution model (170) in respect of at least a portion of the array of sensing pixels (102) of a detector device (100). The array of sensing pixels comprises clusters of pixels. A pixel (140) from the array of sensing pixels is then selected (202) and a first distance and a second distance from the selected pixel to a first neighbouring pixel (142) and a second neighbouring pixel (144), respectively, are determined (402) and the intensity distribution model (170) referenced (406) by the first distance is used to calculate a first weight and a second weight to apply to the first and second neighbouring pixels, respectively. The first distance comprises an intra-cluster distance and the second distance comprises an inter-cluster distance, the intra-cluster distance being different from the inter-cluster distance.

Abstract: A method for reducing jitter of a mouse when in a stationary condition, including: receiving a plurality of raw delta; comparing a movement of the mouse with a predetermined threshold; and when the movement is less than the predetermined threshold, entering a dynamic downshift mode. Entering the dynamic downshift mode includes: storing raw delta into a multi-tap buffer; when the multi-tap buffer is not full, outputting the raw delta as reported delta; and when the multi-tap buffer is full, calculating an average of the total raw delta stored in the multi-tap buffer, and outputting the average delta as reported delta.

Abstract: A method includes capturing, by a camera disposed behind a display panel of an electronic device, an original image through a semi-transparent pixel region of the display panel, and determining a depth position with respect to at least one object identified within the original image. The method further includes accessing, based on the depth position, a plurality of point spread functions (PSFs) corresponding to a plurality of lateral positions at the depth position, and generating a set of image patches based on the plurality of PSFs. Each image patch of the set of image patches is generated based on a different one of the plurality of PSFs. The method concludes with generating a reconstructed image corresponding to the original image based on the set of image patches.

Abstract: A method for mitigating motion blur in a visual-inertial tracking system is described. In one aspect, the method includes accessing a first image generated by an optical sensor of the visual tracking system, accessing a second image generated by the optical sensor of the visual tracking system, the second image following the first image, determining a first motion blur level of the first image, determining a second motion blur level of the second image, identifying a scale change between the first image and the second image, determining a first optimal scale level for the first image based on the first motion blur level and the scale change, and determining a second optimal scale level for the second image based on the second motion blur level and the scale change.

Abstract: An image processor eliminates a character or object from a sequence of frames and then merges the resulting images with those of nearby frames, both preceding and succeeding, to synthesize the background of the sequence of frames.

Abstract: A method of operating an electronic device includes displaying a preview image in response to execution of a camera application, extracting feature information from the preview image, converting user input to an input value in response to the user input generated on the preview image, setting a depth based on the feature information and the input value, and generating a result image in accordance with the depth in response to execution of an imaging operation.

Abstract: Systems and methods are directed to inpainting video. More specifically, initial video data including a sequence of image frames containing missing or corrupted pixel information may be received. Optical flow displacement values and optical flow validity masks may be generated for neighboring image frames of initial video data. Image features from image feature maps of one or more neighboring image frames may be warp-shifted to image feature maps of a current image frame using the optical flow displacement values and warp-shifted image features from the feature maps of the one or more neighboring image frames may be selected based on one or more of the optical flow validity masks. A sequence of complete image frames may be generated based on the selected warp-shifted image features from the feature maps of the one or more neighboring image frames and image features from the image feature maps of the current image frame.

Abstract: A method of encoding a video data signal (15) is provided, together with a method for decoding. The encoding comprises providing color information (51) for pixels in an image, providing a depth map with depth information (52) for the pixels, providing transition information (56, 57, 60, 70, 71) being representative of a width (63, 73) of a transition region (61, 72) in the image, the transition region (61, 72) comprising a depth transition (62) and blended pixels in which colors of a foreground object and a background object are blended, and generating (24) the video data signal (15) comprising encoded data representing the color information (51), the depth map (52) and the transition information (56, 57, 60, 70, 71). The decoding comprises using the transition information (56, 57, 60, 70, 71) for determining the width (63, 73) of the transition regions (61, 72) and for determining alpha values (53) for pixels inside the transition regions (61, 72).

Abstract: A method for deblurring a blurred image includes encoding, by at least one processor, a blurred image at a plurality of stages of encoding to obtain an encoded image at each of the plurality of stages; decoding, by the at least one processor, an encoded image obtained from a final stage of the plurality of stages of encoding by using an encoding feedback from each of the plurality of stages and a machine learning (ML) feedback from at least one ML model; and generating, by the at least one processor, a deblurred image in which at least one portion of the blurred image is deblurred based on a result of the decoding.

Abstract: An image processing device according to the present invention includes: a memory; and at least one processor coupled to the memory. The processor performs operations. The operations includes: detecting a feature point to be used for authentication of a target object included in an image; calculating a first blur being a blur of a predetermined processing point in the feature point; and estimating, by using the first blur, a second blur being a blur of the target object.

Abstract: A system and method for improving a video characteristic of a video stream is described. According to various implementations of the invention, a changed region between a later-in-time image frame and an earlier-in-time image frame and an unchanged region between such two image frames are determined. A new improvement to the video characteristic is determined and applied to the changed region of the later-in-time image frame. A prior improvement to the video characteristic that was determined for the earlier-in-time image frame is applied to the unchanged region of the later-in-time image frame.

Abstract: Systems and methods for implementing array cameras configured to perform super-resolution processing to generate higher resolution super-resolved images using a plurality of captured images and lens stack arrays that can be utilized in array cameras are disclosed. An imaging device in accordance with one embodiment of the invention includes at least one imager array, and each imager in the array comprises a plurality of light sensing elements and a lens stack including at least one lens surface, where the lens stack is configured to form an image on the light sensing elements, control circuitry configured to capture images formed on the light sensing elements of each of the imagers, and a super-resolution processing module configured to generate at least one higher resolution super-resolved image using a plurality of the captured images.

Abstract: Systems and methods are provided for receiving, at a server, a selection of an anchor product from an electronic catalog stored in at least one storage device communicatively coupled to the server, and vectorizing at least one of text and images associated with the selected anchor product and other products in the catalog. At least one of key words may be determined from text data and key images from image data for each product of the catalog. Vectors may be formed from at least one of the keywords and key images, and concatenating the separate vectors together to form final vectors for the products. A similarity search may be performed using the final vectors to determine a group of similar products from the vectorized products of the catalog. Selected products that are within a same slot as the anchor product may be labelled in batch.

Abstract: An image processing method and device, storage medium and electronic device for deblurring an image. The method includes obtaining an image processing instruction including an instruction to deblur a target blurred image; obtaining a target model by training an original model based on a plurality of sample images of different scales, one of the plurality of sample images being a blurred image composed of a plurality of clear images, and the obtained target model being used for deblurring the blurred image to obtain a clear image; based on the image processing instruction, using the target model to deblur the target blurred image to obtain a target clear image; and outputting the target clear image.

Abstract: Embodiments of this application provide an image processing method performed at a computing device. The image processing method includes: obtaining a to-be-processed image with ghost reflection; calculating an image gradient of the to-be-processed image; determining, according to the image gradient, a gradient of a target image obtained after the ghost reflection is removed from the to-be-processed image; and generating the target image based on the gradient of the target image. According to the technical solution of the embodiments of this application, the ghost reflection in the image can be effectively removed, ensuring high quality of a processed image.

Abstract: The present disclosure provides method and system for auto focusing a microscopic imaging system using machine learned regression system such as Convolutional Neural Network (CNN). The method comprises receiving a first image of a sample under review and a second image of sample wherein the first image is captured at first focus position and second image is captured at second focus position. The CNN is trained using the plurality of historic difference images along with direction of focus and optimal focus position. The difference of two images are obtained in terms of difference in pixel values. The direction of focus and optimal focus position for difference image is identified based on plurality of historic difference images along with direction of focus and optimal focus position. The method enables automated stage comprising sample to move towards direction of focus and position at optimal focus position for capturing a focused image.

Abstract: An image processing device acquires a captured image and a depth image. A region dividing section divides the plane of the captured image into regions by determining distance values indicated by the depth image with respect to a preset threshold value (region-divided image). A compressing section generates and outputs data of a captured image having different levels of detail in the respective regions, with respect to the resolution and the number of gradations representing pixel values of the captured image, according to the result of the region division.

Abstract: Systems and methods to provide visual references to passengers in vehicles to prevent motion sickness. The system can include a controller and one or more projectors and/or displays. The controller can detect movement of a vehicle and project images within the vehicle that comport with the detected movement. The system can include a projector to project images on the interior of the vehicle. The system can include one or more displays to display images inside the vehicle. The controller can receive data from one or more cameras, accelerometers, navigation units, magnetometers, and other components to detect the motion of the vehicle. The system can display visual references on the dashboard, door panels, and other interior surfaces to complete the view of passengers, or provide other visual reference, to prevent motion sickness.

Abstract: A vehicle includes one or more cameras that capture a plurality of two-dimensional images of a three-dimensional object. A light detector and/or a semantic classifier search within those images for lights of the three-dimensional object. A vehicle signal detection module fuses information from the light detector and/or the semantic classifier to produce a semantic meaning for the lights. The vehicle can be controlled based on the semantic meaning. Further, the vehicle can include a depth sensor and an object projector. The object projector can determine regions of interest within the two-dimensional images, based on the depth sensor. The light detector and/or the semantic classifier can use these regions of interest to efficiently perform the search for the lights.

Abstract: A system and method for refocusing an image including determining a point spread function (PSF) according to region of interest (ROI) and pixel depth, and converting the PSF to a gradient domain including differentiating the PSF to reduce nonzero elements in the PSF. The technique spreads intensity of pixels into a circle of confusion per the differentiated PSF. A shape of an optical system or aperture of the imaging device may be considered.

Abstract: There is provided an apparatus with a sample holder to hold a sample to be imaged. An image capture device has a field of view and captures an image of the field of view. Also provided is an actuator. A controller controls the actuator to cause relative movement between the sample holder and the image capture device at a given speed and at a given direction during an exposure time of the image capture device such that, in use, the sample moves across at least a portion of the field of view during the exposure time. A processor performs a deblur algorithm to deblur the image using the given speed and the given direction.

Abstract: An imaging device (1a, 1b, 1c, 1d, 1e, 1f) includes at least one imager (Cam0, Cam1, Cam2, Cam3, Cam4, Cam5, Cam6, Cam7) including an imaging element (210, 212) configured to receive light incident through a lens (240); and a casing (10a, 10b, 10c 10d 10e, 10f) at which at least four of the imagers are arranged, the casing being configured such that each one of the imagers and another one of the imagers have optical axes substantially parallel to each other and have opposite incident directions of light on the optical axes, and each one of the imagers is arranged outside imagable ranges of the other imagers.

Abstract: A method and system for determining an optimal wireless channel for a Wi-Fi access point in a cloud-based software defined network (SDN) using an Application Programming Interface (API) is described. Initially, the API collects the measurement information and media access control (MAC) addresses from a plurality of nearby Wi-Fi access points to a client device. The measurement information is analyzed to determine a location of the client device using an Artificial Intelligence (AI) model. To determine the location of the client device, the API creates a plurality of clusters of Wi-Fi access points and updates the plurality of clusters with one or more newly identified Wi-Fi access points. Thereafter, the API calculates an optimal wireless channel for a Wi-Fi access point from the cluster using the AI model and enables the client device to automatically switch to the optimal wireless channel for connecting with the Wi-Fi access point.

Abstract: A sample positioning system is disclosed. The system may include an imaging detector. The system may further include a controller communicatively coupled to the imaging detector and a sample stage, the controller including one or more processors configured to execute program instructions causing the one or more processors to: receive a current position of a field of view of the imaging detector; receive a target position of the sample; receive locations of reference features along a path; direct the sample stage to translate the sample along the path; direct the imaging detector to capture images of the reference features along the path; determine positioning errors of the sample stage along the path based on the images of the reference features; and adjust at least one of a velocity of the sample stage or the path to the target location based on the positioning errors during translation.

Abstract: An object of the present invention is to generate a high-dynamic-range image with fewer pseudo-contours. An apparatus of the present invention generates a high-dynamic-range image and includes a setting unit for setting an imaging condition for each region that constitutes an image. The apparatus includes a setting unit configured to set an exposure condition for the region and a correcting unit configured to correct the set exposure condition so that a difference between a maximum value and a minimum value out of the set exposure condition is small.

Abstract: Devices, methods, and non-transitory program storage devices are disclosed herein to provide improved image processing, the techniques comprising: obtaining an input image and target image data, and then calculating derivatives for the target image data using a regularized derivative kernel operator. In some embodiments, the regularized operator may comprise the following operator: [?1 (1+?)], wherein ? may be a controllable system parameter and preferably is independent of the particular type of image processing being applied to the image. In some embodiments, the techniques may find look-up table (LUT) mappings or analytical functions to approximate the derivative structure of the target image data. Finally, the techniques disclosed herein may generate an output image from the input image based on attempting to closely approximate the calculated derivatives for the target image data. In preferred embodiments, by controlling the mapping, e.g.

Abstract: A method for correcting blur effects is presented. The method includes generating a plurality of images from a camera, synthesizing blurred images from sharp image counterparts to generate training data to train a structure-and-motion-aware convolutional neural network (CNN), and predicting a camera motion and a depth map from a single blurred image by employing the structure-and-motion-aware CNN to remove blurring from the single blurred image.

Abstract: An apparatus executes the instructions to create a combined image by performing combining processing using a plurality of images of an object captured at different focus positions in an optical axis direction, and acquire contrast values of the plurality of images. A depth of field of the combined image is deeper than a depth of field of each of the plurality of images. The at least one processor further executes instructions to calculate a combining ratio used in the combining processing using the contrast values of the plurality of images. An image in which the object is not in focus among the plurality of images is determined using the contrast values of the plurality of images, and the determined image is not used in the combining processing.

Abstract: An imaging device is provided that can correct more accurately an image shake. The imaging device includes an imaging element taking an object image to generate image data, a gyro sensor unit including a gyro sensor, an image-shake correction unit holding the imaging element, moving the imaging element based on an output of the gyro sensor, and correcting a shake of an object image taken by the imaging element, and a mount base holding the gyro sensor unit and the image-shake correction unit.

Abstract: A method of deblurring an observed image acquired by an image sensor in order to determine an observable image corresponding to the deblurred observed image. The observable image and the observed image each are formed by a set of pixels defined by at least one numerical value. The method involves expressing the observable image as a solution minimizing a Tikhonov functional defined by the observed image, applying a spreading function at the point of the image sensor and applying a wavelet transform operator. The Tikhonov functional is expressed as a sum of at least two terms. The method further involves determining the observable image by implementing an iterative projection algorithm released on a closed convex, at each iteration of which, the successive application of a projection operator associated with each of the terms of the Tikhonov functional.

Abstract: A method of sampling and decoding of a signal of interest x comprising, at a training stage: acquiring a set of training signals {xi}i=1M, providing a sampling operator P? and a decoder g?g(.), training operator P? on signals {xi}i=1M to obtain a learned sampling operator P{circumflex over (?)}; and, at a sampling stage: applying P{circumflex over (?)} in a transform domain ? to signal x, resulting in observation signal y; applying the decoder g?g(.) to y, to produce an estimate {circumflex over (x)} of signal x to decode and/or, decide about, the signal. Decoder g?g(.) is trained jointly with P? on signals {xi}i=1M, to obtain a learned decoder g{circumflex over (?)}g, by jointly determining, during a cost minimization step, sampling parameters ? and decoding parameters ?g according to a cost function, and wherein the step of applying the decoder g?g(.) uses decoding parameters ?g, such that estimate {circumflex over (x)} is produced by the learned decoder g{circumflex over (?)}g.

Abstract: System and method of partitioning a plurality of data objects that are each represented by a respective high dimensional feature vector is described, including performing a hashing function on each high dimensional feature vector to generate a respective lower dimensional binary compact feature vector for the data object that is represented by the high dimensional feature vector; performing a further hashing function on each compact feature vector to assign a sub-index ID to the compact feature vector; and partitioning the compact feature vectors into respective partition groups that correspond to the sub-index IDs assigned to the compact feature vectors.

Abstract: Storage allocation based on secure data comparisons is disclosed. One example is a system including a plurality of intermediaries, a data allocator and a plurality of storage containers. Each intermediary receives a request from the data allocator to identify a target storage container of the plurality of storage containers, for secure allocation of a data term. Each intermediary compares, for each storage container, the truncated data term with a collection of truncated candidate terms to select a representative term of the candidate terms, identifies the selected representative term to the storage container, receives a similarity profile from each storage container, where the similarity profile is representative of similarities between the truncated data term and terms in the storage container, and selects a candidate target storage container based on similarity profiles received from each storage container.

Abstract: An image processing system receives an input depth image with a surface that is not developable and generates an output depth image with a piecewise developable surface that approximates the input depth image. Height values for the output depth image are determined using an optimization problem that balances data fidelity and developability. Data fidelity is based on minimizing differences in height values of pixels in the output depth image and height values of pixels in the input depth image. Developability is based on rank minimization of Hessians computed for pixels in the output depth image. In some configurations, the optimization problem is formulated as a semi-definite programming problem and solved using a tailor-made alternating direction method of multipliers algorithm.

Abstract: An imaging apparatus includes a large imaging section, a plurality of small imaging sections, and an image processing section. The plurality of small imaging sections are smaller in optical size than the large imaging section. The large imaging section captures an image of a subject outside the imaging apparatus. The plurality of small imaging sections are provided at positions around the large imaging section to capture images of the subject. The image processing section generates data to be output on the basis of the image captured by the large imaging section and the images captured by the plurality of small imaging sections.

Abstract: A device and algorithm for allowing a customer to choose a photo book cover template that is compatible with a photo having faces. The photo is compared with a set of templates arranged in a first order to determine how compatible the photo is to each of the templates, and a score indicative of compatibility is assigned. A re-sorted set of compatible templates combined with the photo is presented to the customer for consideration.

Abstract: A method is presented for detecting visual artifacts such as macroblocking artifacts appearing in a rendering of a video stream. A decoded video stream that includes a plurality of frames is received. Edges appearing in the plurality of frames are detected and data characterizing the edges is directed to a neural network. The neural network determines a likelihood that a macroblocking artifact occurs in the video stream using the neural network. The neural network is trained using training data that includes (i) one or more video sequences having known macroblocking artifacts whose size and duration are known and (ii) user ratings indicating a likelihood that users visually perceived macroblocking in each of the one or more video sequences when each of the one or more video sequences are rendered.

Abstract: This application discloses a technology for flattening a photographed page of a book and straightening texts therein. The technology uses one or more mathematical models to represent a curved shape of the photographed page with certain parameters. The technology also uses one or more photographic image processing techniques to dewarp the photographed page using the parameters of the curved shape. The technology uses one or more additional parameters that represent certain features of the photographed page to dewarp the photographed page.

Abstract: An approach is provided in which a system transforms a set of embedding approximation values corresponding to a set of knowledge graph nodes into a set of binary valued embedding vectors. The system evaluates the set of binary valued embedding vectors against a query and a selects one of the binary valued embedding vectors based on the evaluation. The system then identifies one of the knowledge graph nodes that corresponds to the selected binary valued embedding vector and in turn, provides a result to the query based on the identified knowledge graph node.

Abstract: An image processing system may include a processor and an associated memory configured to store training data that includes training geospatial images. The processor may be configured to blur each of the training geospatial images. The processor may also be configured to iteratively operate a training model to identify a given feature from each of the blurred training geospatial images so that the blurring is reduced with each iteration.

Abstract: An approach is provided in which a system transforms a set of embedding approximation values corresponding to a set of knowledge graph nodes into a set of binary valued embedding vectors. The system evaluates the set of binary valued embedding vectors against a query and a selects one of the binary valued embedding vectors based on the evaluation. The system then identifies one of the knowledge graph nodes that corresponds to the selected binary valued embedding vector and in turn, provides a result to the query based on the identified knowledge graph node.

Abstract: An embodiment of the disclosure discloses an electronic device including at least one camera module to capture an image based on a specified angle of view, a display to display the image, a processor electrically connected with the camera module and the display. The processor detects a plurality of objects in the image, and performs setting of image data associated with at least one object of the plurality of objects such that a region, which corresponds to the at least one object, in a screen region of the display has a specified light transmittance value. Moreover, various embodiment found through the disclosure are possible.

Abstract: A system configured to provide a three-dimensional representation of a physical environment. The three-dimensional representation including annotation data associated with particular objects and/or viewpoints of the three-dimensional representation. In some cases, the viewpoints may be rendered using image data associated with a photograph captured from a corresponding viewpoint within the physical environment.