Abstract: A computer-implemented method to background noise from a medical image comprising voxels, each voxel having a voxel intensity, the method comprising: generating a mask based on the medical image, wherein the mask comprises a foreground portion designating foreground voxels, a background portion designating background voxels, and a perimeter separating the foreground portion from the background portion; designating a perimeter portion of the mask, the perimeter portion enclosing the perimeter, a subset of the foreground portion, and a subset of the background portion; designating a threshold for the perimeter portion to separate voxels based on voxel intensity; filtering the medical image with the mask and the threshold to obtain a filtered image; and displaying the filtered image on a display.

Abstract: This disclosure provides systems, methods, and devices for image processing that support noise reduction in low-light video sequences. The noise reduction is accomplished through a cascaded set of operations that are configured based on each set of operations location within the cascaded pipeline. The cascade may be implemented as a series of cascaded image post-processing engines (IPEs) within an image signal processor (ISP).

Abstract: Embodiments of the application provide a method and an apparatus for data processing based on event camera. The method includes: obtaining an event stream generated by an event camera of an XR device; performing an optical flow estimation based on the event stream, and determining motion offset information of a first image based on a result of the optical flow estimation, the first image being an image captured by the first camera; performing motion blur compensation on the first image based on the motion offset information to acquire a first compensated image; and obtaining a depth image corresponding to the first image, and determining a binocular video see-through image based on the first compensated image and the depth image. In the method, the motion offset information of a first image acquired by performing an optical flow estimation based on the event stream generated by an event camera is more accurate.

Abstract: An apparatus for identification of a plumbing product includes a communication interface and a controller. The communication interface is configured to receive a raw image of the plumbing product. A first model is configured to analyze the raw image of the plumbing product. A second model is configured to analyze the raw image of the plumbing product in combination with supplemental information for the plumbing product. A third model is configured to analyze a cropped image of the plumbing product. The controller is configured to perform analysis using the models, such that the second model and the third model are performed in series when the first model indicates an object match for the plumbing product in the raw image, and the second model and third model are performed in parallel when the first model lacks the object match for the plumbing product in the raw image.

Abstract: A receptacle analysis and tracking-based time series crop growth analysis method is proposed. The method may include acquiring a crop image representing an image of a crop taken by a camera, and recognizing from the crop image predetermined unit objects included in the crop. The method may also include clustering the predetermined unit objects to form a plurality of clusters, and reconstructing the plurality of clusters on the basis of a receptacle unit. The method may further include generating linkage information between the unit objects and clusters or between the plurality of clusters in each image having different temporal information.

Abstract: A method for cable wire insertion verification includes receiving, from a camera system, an inspection image of a connector housing held in a housing retainer of a housing inspection system. Via a connector recognition machine vision system, the connector housing is classified as a recognized connector housing type based, at least in part, on a comparison between the inspection image of the connector housing and a template image corresponding to the recognized connector housing type. For one or more cable cavities of the connector housing, the method includes automatically verifying whether a correct cable wire is inserted into the cable cavity according to the recognized connector housing type.

Abstract: Disclosed is a motion generation device tagged with text and an operation method thereof. The motion generation device may include a memory configured to store at least one instruction; and at least one processor configured to execute the at least one instruction stored in the memory, wherein the at least one processor is configured to: obtain animation data including a character, convert the obtained animation data into intermediate data for generating a motion of the character over a plurality of frames, generate the motion of the character based on the converted intermediate data, generate a caption for each of the plurality of frames included in the generated motion of the character, generate a text corresponding to the motion of the character by providing the plurality of generated captions to a language model, and generate the text-tagged motion by labeling the generated text to the motion of the character.

Abstract: Systems and methods for generating images using hybrid sampling include obtaining a noisy image and generating a first denoised image during a first reverse diffusion phase using a diffusion neural network. The first denoised image is generated based on a first sampler that uses a first sampling density during at least a portion of the first reverse diffusion phase. Subsequently, a second denoised image is generated based on the first denoised image during a second reverse diffusion phase using the diffusion neural network. The second denoised image is generated based on a second sampler that uses a second sampling density different from the first sampling density during at least a portion of the second reverse diffusion phase.

Abstract: The present disclosure relates to a computer implemented method for detecting artifacts in infrared (IR) illuminated IR images. Artifacts are separated from real objects by analyzing results of regression line fits to histograms of pixel values. Artifacts will have smooth transition gradients and therefore will show smaller deviations of the regression line fit to the histogram data.

Abstract: An inspection system may perform an inspection recipe by receiving reference SHG images of a reference structure based on a scan of the reference structure with an illumination beam and collecting second harmonic generation (SHG) light in response to the illumination beam, where the reference structure includes a multilayer structure including one or more inversion-symmetric materials. The system may further receive test SHG images of a test structure based on a scan of the test structure by illuminating the test structure with the illumination beam and collecting the SHG light in response to the illumination beam, where the test structure and the reference structure have a common design. The system may further identify defects in the test structure by comparing the test and reference SHG images.

Abstract: An embodiment of the present disclosure provides a three-dimensional image generation method, which is performed by a server and is capable of domain adaptation, including generating N target images corresponding to a second domain by converting styles of previously collected N source images corresponding to a first domain according to instructions of an input text, selecting only a target image that satisfies a preset condition among the N target images, and generating multiple three-dimensional images corresponding to a specific domain through certain noise data and a preset camera pose parameter by training a three-dimensional generation model, which is previously built, by using the selected target image.

Abstract: Various implementations disclosed herein include devices, systems, and methods that track a movement of an input device. For example, an example process may include determine a pose of a tracking device in a physical environment based on first sensor data from an image sensor. The process then may receive, from the tracking device, first positional data corresponding to a first relative positioning between the tracking device and an input device in the physical environment, where the first positional data is determined based on second sensor data obtained via a sensor on the tracking device. The process then may track movement of the input device in the physical environment based at least in part on the first positional data and the pose of the tracking device. The process then may determine an input for the electronic device based at least in part on tracking the movement of the input device.

Abstract: Classification of images is inherently a semi-supervised classification problem. Often, the labeled pixels and the unlabeled pixels in the image may have different distribution. Hence classification accuracy of such images is affected. The present disclosure proposes an umbrella framework for semi-supervised learning that considers the domains shifts in labeled and unlabeled pixels. The method proposed a two way optimization solution using deep learning models based on spectral features, spatial features, and fused spectral-spatial features. The model is trained in such a way that it is not only trained on the correct class of pixel but also on the source category of the pixel, for example, labeled pixel or unlabeled pixel. The error in the pixel class is minimized, whereas the error in the source category is encouraged simultaneously.

Abstract: Devices, methods, and non-transitory computer readable media are disclosed herein to repair or mitigate the appearance of unwanted reflection artifacts in captured video image streams. These unwanted reflection artifacts often present themselves as brightly-colored spots that reflect the shape of a bright light source in the captured image. These artifacts, also referred to herein as “green ghosts” (due to often having a greenish tint), are typically located in regions of the captured images where there is not actually a bright light source located in the image. In fact, such unwanted reflection artifacts often present themselves on the image sensor across the principal point of the lens from where the actual bright light source in the captured image is located. Such devices, methods and computer readable media may be configured to detect, track, and repair such unwanted reflection artifacts in an intelligent fashion, e.g., leveraging images captured with varying exposure settings.

Abstract: A method, system, and article is directed to automatic content-dependent image processing algorithm selection.

Abstract: This system for training an operation determination model for a medical instrument control device generates training data by means of a reinforcement learning model, and, using the training data, can train an operation determination model configured to output information associated with operational commands for a driving unit that transports medical instruments.

Abstract: A document creation support apparatus comprising at least one processor, wherein the processor is configured to: extracts regions having one or more preset physical features from a medical image, and generates a comment on findings using a disease name associated with a physical feature of at least one of the extracted regions.

Abstract: Embodiments of the present application provide an image enhancement method, a chip, and an image acquisition device. The method comprises: storing, by a main processor, a preset number of Raw image frames to be processed currently acquired from a sensor in a first off-chip memory, into the on-chip memory; performing, by a first neural network processor, image fusion on a preset number of Raw image frames to be fused currently obtained based on the preset number of Raw image frames to be processed currently in the on-chip memory based on a preset image fusion network, to obtain a fused Raw image frame, and storing it into the on-chip memory; sending, by the main processor to an ISP chip, a target Raw image frame to be sent currently obtained based on the fused Raw image frame in the on-chip memory.

Abstract: Systems, methods, models, and training data for models are discussed, for determining vehicle positioning, and in particular identifying tailgating. Simulated training images showing vehicles following other vehicles, under various conditions, are generated using a virtual environment. Models are trained to determine following distance between two vehicles. Trained models are used to in detection of tailgating, based on determined distance between two vehicles. Results of tailgating are output to warn a driver, or to provide a report on driver behavior.

Abstract: Methods, systems, and computer readable media are disclosed for characterizing the porosity of a microspheric material, training a classifier model for classifying spatially-contiguous image regions of backscatter electron scanning electron microscopy (BSE-SEM) images of a microspheric material according to the particle composition of the image regions, and characterizing the elemental composition of a microspheric material. The methods include: receiving image data representative of a microscopic images of a sample microspheric material; segmenting the image data into a plurality of spatially-contiguous image regions; classifying, by a trained machine-learning model, each of the plurality of spatially-contiguous image regions; and characterizing the microspheric material. The disclosure also relates to compositions for use as an FCC catalyst comprising microspheres which comprise alumina and/or clay and/or a zeolite.