An image processing apparatus counts at least one of the number of pixels having an identical color to a target pixel, the number of pixels having a similar color to the target pixel, and the number of pixels having a different color from the target pixel in a target window, and determines an attribute of the target pixel based on a result of the counting.

Training a generative adversarial network (GAN) for use in facial recognition, comprising providing an input image of a particular face into a facial recognition system to obtain a faceprint; obtaining, based on the input faceprint and a noise value, a set of output images from a GAN generator; obtaining feedback from a GAN discriminator, wherein obtaining feedback comprises inputting each output image into the GAN discriminator and determining a set of likelihood values indicative of whether each output image comprises a facial image; determining, based on each output image, a modified noise value; inputting each output image into a second facial recognition network to determine a set of modified faceprints; defining, based on each modified noise value and modified faceprint, feedback for the GAN generator, wherein the feedback comprises a first value and a second value; and modifying control parameters of the GAN generator.

A method and apparatus for classifying light detection and ranging sensor data are provided. The method includes transforming sensor data of the LIDAR into point cloud data, selecting a cell including a subset of the point cloud data, dividing the selected cell into a plurality of voxels, calculating a difference of gradients for the plurality of voxels, performing a first pass on the plurality of voxels to identify voxels that contain an object based the difference of gradients, performing a second pass on the plurality of voxels to identify voxels that contain the object by adjusting a voxel with at least one from among a jitter parameter and a rotation parameter, and outputting a centroid average of voxels identified as containing the object.

An image processing method includes obtaining a sensed image, wherein the sensed image comprises a pattern; dividing the sensed image into a plurality of blocks; calculating a direction field according to the pattern in each of the blocks; calculating a similarity degree between the direction field of a first block and the direction fields of adjacent blocks of the first block; and classifying the first block into a first part according to the similarity degree of the first block.

With regard to an inspection region for inspecting abnormality of a holding state of the substrate in an image of the substrate holding unit, (1) an upper end surface of the substrate being normally held by the substrate holding unit is confirmed, (2) based on a position of the upper end surface of the substrate that has been confirmed, a position of the inspection region in a vertical direction is determined, and (3) for a candidate of the inspection region of which the position in the vertical direction has been determined, density thereof at a rotation start time of the substrate holding unit is obtained, a horizontal position of the inspection region is determined based on a difference image integrated value, which is an integrated value of a difference absolute value with density of the same region in an initial state of the substrate holding unit.

In a system for determining liveness of an image presented for authentication, a reference signal is rendered on a display, and a reflection of the rendered signal from a target is analyzed to determine liveness thereof. The analysis includes spatially and/or temporally band pass filtering the reflected signal, and determining RGB values for each frame in the reflected signal and/or each pixel in one or more frames of the reflected signal. Frame level and/or pixel-by-pixel correlations between the determined RGB values and the rendered signal are computed, and a determination of whether an image presented is live or fake is made using either or both correlations.

Various embodiments provide an optimized image filter. The optimized image and video obtains an input image and selects a target pixel for modification. Difference values are then determined between the selected target pixel and each reference pixel of a search area. Subsequently, a weighting function is used to determine weight values for each of the reference pixels of the search area based on their respective difference value. The selected target pixel is then modified by the optimized image filter using the determined weight values. A new target pixel in an apply patch is then selected for modification. The new target pixel is modified using the previously determined weight values reassigned to a new set of reference pixels. The previously determined weight values are reassigned to the new set of reference pixels based on each of the new set of reference pixels' position relative to the new target pixel.

A method includes determining, at an image processing device, object quality values for a plurality of objects based on portions of image data corresponding to an image. Each portion corresponds to an object of the plurality of objects represented in the image. The method includes accessing, via the image processing device, object category metrics associated with an object category corresponding to each object of the plurality of objects. The method also includes determining, with the image processing device, image processing for the image based on comparisons of the object quality values for each object to corresponding object category metrics.

A CT system includes a gantry having a rotatable base and having an opening for receiving an object to be scanned, an x-ray source, a CT detector, and a computer programmed to detect a mis-registration at a slab boundary between a first slab and a second slab of a reconstructed image, quantify an amount of mis-registration at the slab boundary, and adjust the reconstructed image at the slab boundary based on the quantification.

Systems and methods identify consumer products in images. Known consumer products are captured as grayscale or color images. They are converted to binary at varying thresholds. Connected components in the binary images identify image features according to pixels of a predetermined size, shape, solidity, aspect ratio, and the like. The image features are stored and searched for amongst image features similarly extracted from unknown images of consumer products. Identifying correspondence between the features of the images lends itself to identifying or not known consumer products.