Abstract: An information processing method includes processing of: acquiring, from a plurality of time-series images in which an object is captured, first information including at least a plurality of positions or a plurality of sizes of the object; executing prediction processing of predicting second information including at least one of a position or a size of the object at a next time point in a time-series based on the first information and recursively executing the prediction processing based on the first information and the second information to predict the second information of the object at a time point further next to the next time point; executing recognition processing of recognizing motion of the object based on the second information; and determining a total number of times of recursion of the prediction processing based on a result of the recognition processing.

Abstract: Systems and methods are configured to generate a frequency map representing a density of objects found in regions of a sample that may be used in setting parameters for imaging the regions. Various embodiments involve binarizing the pixels for a raw image of the sample to transform the image into binary data. Run-length encoded components are identified from the data for dimensions of the raw image. Each component is a length of a sequence of adjacent pixels found in a dimension with the same value in the binary data. A projection of the image is then generated from projection values for the dimensions. Each projection value provides a measure of the density of objects present in a dimension with respect to the components identified for the dimension. This projection is used to identify a level of density for each region of the sample from which the frequency map is generated.

Abstract: The present disclosure relates to a method and a system for imaging a biological tissue. A monochromatic image of the biological tissue is obtained. A texture analysis of the biological tissue is performed using spatial information of the monochromatic image to identity features of the biological tissue. A texture image is generated based on the features of the biological tissue. The biological tissue of the subject is classified as normal or abnormal at least in part based on a comparison between first order statistics of the texture image and predetermined values.

Abstract: A system mitigates transitions among pixels a mosaic image. The system uses two balancing adjustments and an easing transition adjustment. The first balancing adjustment uses a histogram to remap the pixels in the mosaic to a normal distribution. The second balancing adjustment minimizes transitions between adjacent individual images in the mosaic based on pixels positions of edge pixels and intensity differences between edge pixels in the adjacent images. The easing transitions adjustment modifies the intensity of a target pixel as a function of a radius from the target pixel, the distance between the target pixel and a nearest neighbor pixel, and an intensity difference between the nearest neighbor pixel and a pixel adjacent to the nearest neighbor pixel. A revised mosaic image is created as a function of the first balancing adjustment, the second balancing adjustment, and the easing transition adjustment.

Abstract: A foveated down sampling and correction (FDS-C) circuit for combined down sampling and correction of chromatic aberrations in images. The FDS-C circuit performs down sampling and interpolation of pixel values of a first subset of pixels of a color in a raw image using down sampling scale factors and first interpolation coefficients to generate first corrected pixel values for pixels of the color in a first corrected version of the raw image. The FDS-C circuit further performs interpolation of pixel values of a second subset of the pixels in the first corrected version using second interpolation coefficients to generate second corrected pixel values for pixels of the color in a second corrected version of the raw image. Pixels in the first subset are arranged in a first direction, pixels in the second subset are arranged in a second direction, and the down sampling scale factors vary along the first direction.

Abstract: Provided is a decoding apparatus including: a communication interface configured to receive AI encoding data generated as a result of artificial intelligence (AI) down-scaling and first encoding of an original image; a processor configured to divide the AI encoding data into image data and AI data; and an input/output (I/O) device, wherein the processor is further configured to: obtain a second image by performing first decoding on a first image obtained by performing AI down-scaling on the original image, based on the image data; and control the I/O device to transmit the second image and the AI data to an external apparatus. In some embodiments, the external apparatus performs an AI upscaling of the second image using the AI data, and displays the resulting third image.

Abstract: The invention discloses an image detection method, apparatus, device and a medium. The method includes: determining one or more edge points in an input image and gradient directions thereof; generating an initial matrix with the same size as the input image, and assigning the same initial value to all matrix elements in the initial matrix; for each pixel point located in an accumulation region of each edge point, assigning a corresponding accumulation value to a matrix element in the initial matrix corresponding to the pixel point, to obtain an accumulation matrix; determining one or more circle-center positions in the input image based on the accumulation matrix; wherein, the accumulation region of each edge point includes a first direction line for accumulation along the gradient direction of the edge point and a second direction line for accumulation along the direction opposite to the gradient direction of the edge point.

Abstract: Various embodiments of an apparatus, methods, systems and computer program products described herein are directed to an agricultural observation and treatment system and method of operation. The agricultural treatment system may determine a first real-world geo-spatial location of the treatment system. The system can receive captured images depicting real-world agricultural objects of a geographic scene. The system can associate captured images with the determined geo-spatial location of the treatment system. The treatment system can identify, from a group of mapped and indexed images, images having a second real-word geo-spatial location that is proximate with the first real-world geo-spatial location. The treatment system can compare at least a portion of the identified images with at least a portion of the captured images. The treatment system can determine a target object and emit a fluid projectile at the target object using a treatment device.

Abstract: Systems and methods for optimal electron beam metrology guidance are disclosed. According to certain embodiments, the method may include receiving an acquired image of a sample, determining a set of image parameters based on an analysis of the acquired image, determining a set of model parameters based on the set of image parameters, generating a set of simulated images based on the set of model parameters. The method may further comprise performing measurement of critical dimensions on the set of simulated images and comparing critical dimension measurements with the set of model parameters to provide a set of guidance parameters based on comparison of information from the set of simulated images and the set of model parameters. The method may further comprise receiving auxiliary information associated with target parameters including critical dimension uniformity.

Abstract: A gaze estimation method and apparatus is disclosed. The gaze estimation method includes obtaining an image including an eye region of a user, extracting, from the obtained image, a first feature of data, obtaining a second feature of data used for calibration of a neural network model, and estimating a gaze of the user using the first feature and the second feature.

Abstract: Systems and techniques for the optimized matched filter tracking of space objects are provided. In one aspect, a method of detecting faint objects includes receiving a plurality of telescope images from a telescope imaging system and performing a plurality of computational calculations on the telescope images using a plurality of search parameters to identify one or more objects moving through the telescope images. The number of the computational calculations is reduced by a priori relatively restricting a parameter space of the computational calculations.

Abstract: Example PPE fit systems are described that verify that a current fit of a PPE on a wearer is consistent with a previous fit test of the PPE on the particular wearer. During a fit test of a PPE, an image capture device captures an image of the wearer with the PPE that represents a verified reference for a fit of PPE of a same or similar type on the particular wearer. During a verification check, an image capture device captures an image of the wearer with the PPE. Visual features of the fit test image are compared with visual features of the verification image to determine whether the fit of the PPE on the wearer is correct. Differences between the visual features are alerted to the user, such as for correction, adjustment, or retest.

Abstract: Methods and systems for image deblurring are described. The weights of a deblurring network are first trained by meta-training the deblurring network on both a primary deblurring task and an auxiliary reconstruction task. Application-time training of the deblurring network is then performed, using an application-time blurry input image, to obtain values of application-time trained weights. Application-time training includes performing the auxiliary reconstruction task on the application-time blurry input image, and updating the weights of the deblurring network based on an auxiliary loss computed from the auxiliary reconstruction task. A deblurred output image is generated from the application-time blurry input image, using the application-time trained weights in the deblurring network.

Abstract: A method for keypoint matching includes determining a first set of keypoints corresponding to a current environment of the agent. The method further includes determining a second set of keypoints from a pre-built map of the current environment. The method still further includes identifying matching pairs of keypoints from the first set of keypoints and the second set of keypoints based on geometrical similarities between respective keypoints of the first set of keypoints and the second set of keypoints. The method also includes determining a current location of the agent based on the identified matching pairs of keypoints. The method further includes controlling an action of the agent based on the current location.

Abstract: Methods and apparatus for shared image processing among multiple devices. In one embodiment, an exemplary action camera performs a partial multiband blend. Even though the action camera may not have resources to handle the multiband blend of the entire action camera's footage, it can do a significant portion. The partially blended content can be used in ready-to-share applications, or completely blended by another device.

Abstract: A video decoding method comprises warping different viewpoint images into a virtual viewpoint by using depth information of an input viewpoint; determining a first weight for the pixels based on quality and depth of pixels of the warped viewpoint images; determining a second weight for the pixels based on a location relation between the pixels of the warped viewpoint images and the virtual viewpoint; determining a blending weight for the pixels based on the first weight and the second weight; blending the warped viewpoint images according to the blending weight; and synthesizing an image of the virtual viewpoint based on the blended viewpoint images.

Abstract: Techniques related to generating a fine super resolution image from a low resolution image including a person wearing a predetermined uniform are discussed. Such techniques include applying a pretrained convolutional neural network to a stacked image including image channels from a coarse super resolution image, label data corresponding to the coarse super resolution image from available labels relevant to the uniform, and pose data corresponding to the person to determine the fine super resolution image.

Abstract: Aspects of the current disclosure include systems and methods for identifying an entity in a query image by comparing the query image with digital images in a database. In one or more embodiments, a query feature may be extracted from the query image and a set of candidate features may be extracted from a set of images in the database. In one or more embodiments, the distances between the query feature and the candidate features are calculated. A feature, which includes a set of shortest distances among the calculated distances and a distribution of the set of shortest distances, may be generated. In one or more embodiments, the feature is input to a trained model to determine whether the entity in the query image is the same entity associated with one of the set of shortest distances.

Abstract: According to one embodiment, a position and attitude estimation device includes a processor. The processor is configured to acquire time-series images continuously captured by a capture device installed on a mobile object, estimate first position and attitude of the mobile object based on the acquired time-series images, estimate a distance to a subject included in the acquired time-series images and correct the estimated first position and attitude to a second position and attitude based on an actual scale, based on the estimated distance.

Abstract: A first six degree-of-freedom (DoF) pose of an object from a perspective of a first image sensor is determined with a neural network. A second six DoF pose of the object from a perspective of a second image sensor is determined with the neural network. A pose offset between the first and second six DoF poses is determined. A first projection offset is determined for a first two-dimensional (2D) bounding box generated from the first six DoF pose. A second projection offset is determined for a second 2D bounding box generated from the second six DoF pose. A total offset is determined by combining the pose offset, the first projection offset, and the second projection offset. Parameters of a loss function are updated based on the total offset. The updated parameters are provided to the neural network to obtain an updated total offset.