Abstract: A method of classification of a film non-uniformity on a substrate includes obtaining a color image of a substrate with the color image comprising a plurality of color channels, obtaining a standard color for the color image of the substrate, for each respective pixel along a path in the color image determining a difference vector between the a color of the respective pixel and the standard color to generate a sequence of difference vectors, sorting the pixels along the path into a plurality of regions including at least one normal region and at least one abnormal region based on the sequence of difference vectors, and classifying the at least one abnormal region as overpolished or underpolished based on at least one difference vector of a pixel at a boundary between the abnormal region and an adjacent normal region.

Abstract: An information acquisition unit acquires a learning image including a disease region and a first teacher label that specifies the disease region included in the learning image. A teacher label generation unit generates at least one second teacher label of which a criterion for specifying the disease region is different from the first teacher label. A learning unit trains a discriminator that detects a disease region included in a target image on the basis of the learning image, the first teacher label, and the at least one second teacher label.

Abstract: An imaging method is described for generating image data of an examination region of an object that is to be examined. First X-ray projection measurement data of the examination region is acquired using a first X-ray energy spectrum and at least second X-ray projection measurement data of the examination region is acquired using a second X-ray energy spectrum which is different from the first X-ray energy spectrum. A priori image data is reconstructed based on at least the first X-ray projection measurement data and a location-dependent distribution of X-ray attenuation values. A basis material decomposition is performed based on the first X-ray projection measurement data and the at least second X-ray projection measurement data. A location-dependent weighting of the basis materials is determined as a function of the location-dependent distribution of the X-ray attenuation values. An image for the examination region is determined by reconstructing virtual basis-material-weighted image data.

Abstract: A method for compressing images based on joint photographic experts group (JPEG) standard includes: compressing data of one or more first image blocks with a first compression level to produce compression data of the one or more first image blocks; adjusting the first compression level to obtain a second compression level according to at least one of a data size-related index regarding the compression data of the one or more first image blocks or a transmission-related index regarding transmission of the compression data of the one or more first image blocks; and compressing data of a second image block with the second compression level.

Abstract: A neural network model training method and an apparatus for complex characteristic classification and common localization are proposed. In the method, a neural network model includes: a convolution layer for performing a convolution operation on an input image by using a convolution filter; a pooling layer for performing pooling on an output of the convolution layer; and class-specific fully connected layers respectively corresponding to classes into which complex characteristics are classified and outputting values obtained by multiplying an output of the pooling layer by class-specific weights (Wfc(Tt). The method includes: (a) inputting the input image to the convolution layer; (b) calculating class-specific observation maps for respective classes on the basis of the output of the convolution layer; (c) calculating an observation loss (Lobs) common to the classes on the basis of the class-specific observation maps; and (d) back-propagating a loss based on the observation loss to the neural network model.

Abstract: A method for detecting defects in images, is employed in a computer device, and stored in a storage medium. The method trains an autoencoder model using unblemished images, inputting an image to be detected into the autoencoder model, and obtaining a reconstructed image. An image error is calculated between the image to be detected and the reconstructed image, and the image error is inputted into a student's t-distribution and a calculation result is obtained. In response that the calculation result falls within a preset defect determination criterion range, the image to be detected is determined to be an unblemished image. In response that the calculation result does not fall within the preset defect determination criterion range, the image to be detected is determined to be a defective image. The method improves the efficiency and accuracy of defect detection.

Abstract: A method for detecting an exit lane for a motor vehicle includes: detecting, in images acquired by an onboard camera, marking lines on the ground on the right and on the left laterally delimiting the main traffic lane, determining first and second heading angles with respect to the detected marking lines, providing a first, standard exit lane detection state, based on the comparison of the first and second heading angles with a predetermined threshold, and providing a second, coaxial exit lane detection state, based on the comparison between the derivatives with respect to time of the variation of the first and second heading angles and of the verification that the value of the first or the second heading angle remains substantially close to a zero value.

Abstract: An image processing apparatus acquires a first low-energy image and a first high-energy image captured in a state in which a grid for removing scattered rays is inserted between a radiation source and a radiation detector, and the radiation source is disposed at a first position at which an incidence direction of radiation is a normal direction with respect to the grid, acquires a second low-energy image and a second high-energy image captured in a state in which the grid retreats from between the radiation source and the radiation detector, and the radiation source is disposed at a second position, estimates a thickness of the subject from the first low-energy image and the first high-energy image, and generates a difference image indicating a difference between the second high-energy image and the second low-energy image from the second low-energy image and the second high-energy image based on the thickness.

Abstract: A method and apparatus for tracking hand joints are disclosed, where the method of tracking hand joints includes estimating angle information of finger joints based on initial positions of hand joints obtained from an image, generating a kinematic model of the hand joints based on the initial positions of the hand joints and the angle information of the finger joints, and tracking the hand joints by updating at least one of positions or angles of the hand joints based on the initial positions of the hand joints and the kinematic model.

Abstract: An automated inventory monitoring system includes an image capture module able to create an image of an aisle of a retail store. Product images and shelf label and peg label images are identified in the image and products are associated with product labels based on the positioning of the products with respect to the labels. Based on the association between labels and products, out-of-stock products are detected and reported to the retail store.

Abstract: Free space machine interface and control can be facilitated by predictive entities useful in interpreting a control object's position and/or motion (including objects having one or more articulating members, i.e., humans and/or animals and/or machines). Predictive entities can be driven using motion information captured using image information or the equivalents. Predictive information can be improved applying techniques for correlating with information from observations.

Abstract: A sensor and/or system controller may process an image multiple times at multiple resolutions to detect glare conditions. A glare condition threshold used to determine whether a glare condition exists may be based on the resolution of the image. When the resolution of the image is higher, the glare condition threshold may be higher. The sensor and/or system controller may organize one or more adjacent pixels having similar intensities into pixel groups. The pixel groups may vary in size and/or shape. The sensor and/or system controller may determine a representative group luminance for the pixel group (e.g., an average luminance of the pixels in the group). The sensor and/or system controller may determine a group glare condition threshold, which may be used to determine whether a glare condition exists for the group of pixels and/or may be based on the size of the group.

Abstract: An image processing apparatus includes a first extraction unit configured to subject an input image to first extraction processing to extract a first feature quantity, a second extraction unit configured to subject the input image to second extraction processing having a lower throughput than the first extraction processing to extract a second feature quantity, a comparison unit configured to compare the first feature quantity with the second feature quantity and then output a comparison result, and an update unit configured to update parameters to be used for the second extraction processing by the second extraction unit, based on the comparison result.

Abstract: An individual identification system includes an acquiring means and a determining means. The acquiring means is configured to acquire a matched image obtained by shooting part of a predetermined region of a matching target object. The determining means is configured to calculate a score representing a degree to which a partial image similar to the matched image exists in a registration image obtained by shooting a predetermined region of a registration target object, and determine based on the score whether or not the matching target object is identical to the registration target object.

Abstract: Described are platforms, systems, and methods for screening patients. In one aspect, a computer-implemented method comprises: receiving, from a cellular imaging device, image data comprising calcium kinetic features of neuronal cultures derived from a patient; processing the image data through a machine-learning model to determine a diagnosis for the patient based on the calcium kinetic features, the machine-learning model trained using neuronal calcium data; and providing the diagnosis a user interface.

Abstract: A method for 3D object detection is described. The method includes predicting, using a trained monocular depth network, an estimated monocular input depth map of a monocular image of a video stream and an estimated depth uncertainty map associated with the estimated monocular input depth map. The method also includes feeding back a depth uncertainty regression loss associated with the estimated monocular input depth map during training of the trained monocular depth network to update the estimated monocular input depth map. The method further includes detecting 3D objects from a 3D point cloud computed from the estimated monocular input depth map based on seed positions selected from the 3D point cloud and the estimated depth uncertainty map. The method also includes selecting 3D bounding boxes of the 3D objects detected from the 3D point cloud based on the seed positions and an aggregated depth uncertainty.

Abstract: A method of processing an immersive video includes classifying view images into a basic image and an additional image, performing pruning with respect to view images by referring to a result of classification, generating atlases based on a result of pruning, generating a merged atlas by merging the atlases into one atlas, and generating configuration information of the merged atlas.

Abstract: Techniques are disclosed for systems and methods for water non-water segmentation of navigational imagery to assist in the autonomous navigation of mobile structures. An imagery based navigation system includes a logic device configured to communicate with an imaging module coupled to a mobile structure and/or configured to capture images of an environment about the mobile structure. The logic device may be configured to receive at least one image from the imaging module; determine a water/non-water segmented image based, at least in part, on the received at least one image, and generate a range chart corresponding to the environment about the mobile structure based, at least in part, on the determined water/non-water segmented image and/or the received at least one image.

Abstract: A method and system are provided for illuminating and imaging a biological sample using a brightfield microscope for the purpose of counting biological cells. The method comprises positioning a sample to be viewed by way of an objective lens of the microscope, the sample comprising a plurality of biological cells; capturing and storing, using an image capturing apparatus, one or more focal image stacks; processing the one or more focal image stacks using a cell localisation neural network, the cell localisation neural network outputting a list of one or more cell locations; determining, using the list of cell locations, one or more cell focal image stacks, each cell focal image stack being obtained from the one or more focal image stacks; processing the one or more cell focal image stacks using an encoder neural network; determining, using the list of cell locations and the list of cell fingerprints, a number of cells within the sample.

Abstract: Systems and methods for computer-implemented pre-optimization of input data before further processing thereof by a computer-implemented analyzation process, such as optical character recognition (OCR). A cooperative model is employed that combines one or more supervised-learning based inspector sub-models, and one or more filter sub-models that operating in series with the inspector sub-model(s). The inspectors first receive the input data and calculate one predicted transformation parameters then used to perform transformations on the input data. The inspector-transformed data is then passed to the filters, which derive respective convolution kernels and apply same to the inspector-transformed data before passing same to the OCR or other analyzation process. The inspectors may be pretrained with different training data.