An image comparison method constituted of: receiving images; point matching the images to identify tentatively corresponding points; responsive to the identified points, defining a plurality of first image tiles within the first image; defining a plurality of second image tiles within the second image, each corresponding to a respective first image tile; adjusting the intensity of a set of pixels responsive to the cumulative relative frequency of the respective pixel intensity within the respective tile; for each tile, applying a plurality of non-linear functions; for each function, separately determining a moment of the outcome of the respective non-linear function for each axis of the respective tile; for each first image tile and corresponding second image tile, determining a distance between the subspaces spanned by moment vectors; and responsive to determined distances, determining whether respective first and second image tiles comprise observations of the same portion.

Predicting patch displacement maps using a neural network is described. Initially, a digital image on which an image editing operation is to be performed is provided as input to a patch matcher having an offset prediction neural network. From this image and based on the image editing operation for which this network is trained, the offset prediction neural network generates an offset prediction formed as a displacement map, which has offset vectors that represent a displacement of pixels of the digital image to different locations for performing the image editing operation. Pixel values of the digital image are copied to the image pixels affected by the operation.

A biometric analysis structure, method and neural network with coded mask are provided. The biometric analysis structure includes a display panel, a light source, and a sensor. The sensor is disposed on the optical path of light from the light source and reflected by the display panel. The biometric analysis structure includes an coded mask. The coded mask is disposed on the optical path in front of the sensor. the coded mask is represented as a first matrix in a matrix form and the first matrix is a delta function after satisfying an autocorrelation operation. The resulting image can be inversely resolved based on the image of the coded mask. Thus, the security of the fingerprint recognition method is improved, and the thickness of the entire imaging structure is reduced.

Predicting patch displacement maps using a neural network is described. Initially, a digital image on which an image editing operation is to be performed is provided as input to a patch matcher having an offset prediction neural network. From this image and based on the image editing operation for which this network is trained, the offset prediction neural network generates an offset prediction formed as a displacement map, which has offset vectors that represent a displacement of pixels of the digital image to different locations for performing the image editing operation. Pixel values of the digital image are copied to the image pixels affected by the operation by: determining the vectors pixels that correspond to the image pixels affected by the image editing operation and mapping the pixel values of the image pixels represented by the determined offset vectors to the affected pixels. According to this mapping, the pixel values of the affected pixels are set, effective to perform the image editing operation.

Provided are an image processing method device, and a storage medium. The method includes: at an electronic device with a processor, a memory and a camera: acquiring first video data for commodities on a shelf; converting the first video data into a plurality of single-frame images; generating a virtual image scenario according to the plurality of single-frame images; acquiring second video data for the commodities on the shelf; performing unit identification on the plurality of single-frame images in the second video data with a preset unit identification model to identify a plurality of commodity regions; labeling the identified plurality of commodity regions; and replacing corresponding regions in the virtual image scenario with ones of the single-frame images with the labeled commodity regions.

A processing system having at least one processor may obtain at least a first source video from a first endpoint device and a second source video from a second endpoint device, where each of the first source video and the second source video is a two-dimensional video, determine that the first source video and the second source video share at least one feature that is the same for both the first source video and the second source video, and generate a volumetric video from the first source video and the second source video, where the volumetric video comprises a photogrammetric combination of the first source video and the second source video.

A detection apparatus to extract features from an image; determine the number of candidate regions of the object in the image based on the extracted features, wherein the determined number of the candidate regions is decided by a position and shape of the candidate regions; and to detect the object from the image based on at least the extracted features and the determined number, position and shape of the candidate regions.

Embodiments include methods, systems, and computer program products for remediating a color vision deficiency. Aspects include receiving a time dependent location information for a user. Aspects also include receiving images of a plurality of objects, wherein each of the plurality of objects corresponds to the time dependent location information and, for each of the plurality of objects, identifying an object type and an object color. Aspects also include determining a number of distinguishable colors required to remediate a color vision deficiency and a number of available colors and overlaying one of the plurality of objects with an available color responsive to a determination that the number of distinguishable colors does not exceed the number of available colors.

Systems and methods for image processing, and specifically for generating object-agnostic image representations, are described. Embodiments of the present disclosure receive a training image including a foreground object and a background, remove the foreground object from the training image to obtain a modified training image, inpaint a portion of the modified training image corresponding to the foreground object to obtain an inpainted training image, encode the training image and the inpainted training image using a machine learning model to obtain an encoded training image and an encoded inpainted training image, and update parameters of the machine learning model based on the encoded training image and the encoded inpainted training image.

Techniques for cloud forecast using configuration engine, segment cloud coverage determination engine, and trained forecast models are described herein. The disclosed techniques include generating a configuration mask based on a type of an optical component for capturing a plurality of sky images and determining cloud coverages of a plurality of corresponding segments by applying the configuration mask and a segment cloud coverage determination algorithm to the plurality of images. The disclosed techniques include training a plurality of forecast models for forecasting cloud coverages of the plurality of corresponding segments using the determined cloud coverages. The plurality of forecast models correlate cloud coverages of the plurality of corresponding segments at a time point with a cloud coverage of any particular segment at a later time point. The plurality of trained forecast models generate data indicative of cloud coverages of the plurality corresponding segments at a future time point.