Abstract: The present disclosure provides a system and method for image data acquisition. The method may include obtaining image data of a subject including a first type of tissue and a second type of tissue. The method may include determining, based on the image data of the subject, a target portion including at least a portion of at least one of the first type of tissue or the second type of tissue. The method may include determining, based at least in part on the target portion represented in the image data, a scan mode corresponding to the target portion. The method may include causing an imaging device to acquire, based on the scan mode, image data of the target portion.

Abstract: A road sign interpretation system includes a front-facing camera mounted on or in a vehicle collecting image data of multiple road signs. A first convolutional neural network (CNN) receives the image data from the front-facing camera and yields a set of sign predictions including one or more sign text instances. A second CNN defining a text extractor receives the image data from the front-facing camera and extracts text candidates including the multiple sign text instances. Sign and sign data localization is provided in the second CNN to compute a text order from the multiple sign text instances. A sign text synthesizer module receives individual sign text instances from the first CNN and individual ones of the sign text instances in digitized forms from an optical character recognizer (OCR). A semantic encoding and interpretation module receives the sign text instances and identifies semantics of the multiple road signs.

Abstract: Automated image analysis used in vascular state modeling. Coronary vasculature in particular is modeled in some embodiments. Methods of “virtual revascularization” of a presently stenotic vasculature are described; useful, for example, as a reference in disease state determinations. Structure and uses of a model which relates records comprising acquired images or other structured data to a vascular tree representation are described.

Abstract: Disclosed herein are methods and systems for detection and discrimination of optical signals from a densely packed substrate. These have broad applications for biomolecule detection near or below the diffraction limit of optical systems, including in improving the efficiency and accuracy of polynucleotide sequencing applications.

Abstract: An image signal processor includes a statistic data generating unit for receiving an image signal from an external device, an image processing unit for receiving the image signal, and a direct memory access (DMA) module connected to the statistic data generating unit and the image processing unit. The statistic data generating unit performs first image pre-processing on the image signal and generates first statistic data based on the image signal subjected to the first image pre-processing. The DMA module stores the first statistic data therein and provides the stored first statistic data to the image processing unit. The image processing unit performs second image pre-processing on the image signal and performs image processing on the image signal based on the first statistic data.

Abstract: Systems, methods, and other embodiments described herein relate to detecting reflection artifacts in a point cloud. In one embodiment, a method includes generating a first convex hull for defining an outline of a first cluster of points in a depth image and generating a second convex hull for defining an outline of a second cluster of points in the depth image. The first cluster of points originates from a point cloud and the second cluster of points originates from the point cloud. The method includes determining whether the second cluster of points is a reflection artifact based on identifying if a portion of the second convex hull is enclosed in the first convex hull and a range of the second convex hull is larger than a range of the first convex hull.

Abstract: The present technology pertains to validating a virtual camera in a simulation environment by utilizing an improved camera model to provide quantitative measurements of the model's performance. A method of validating a virtual camera in a simulation environment comprises presenting at least one reference chart in the simulated environment and capturing images of the reference chart in the simulated environment using a virtual camera. The method further includes interrupting an image pipeline of the virtual camera after at least one simulated process in the image pipeline to extract a RAW image. The method analyzes the RAW image to derive measurements of metrics to characterize the virtual camera. The measured metrics are compared to metrics of a calibrated real-world camera to verify that the metrics are within a threshold delta that is indicative that the virtual camera sufficiently approximates the real-world camera.

Abstract: A method of training a generative adversarial network for performing semantic segmentation of images. The generative adversarial network includes a generator neural network and a critic neural network. The method includes using the generator neural network to generate predicted image segmentation maps from input images, wherein each predicted image segmentation map includes a classification prediction for each of a plurality of pixels of a respective input image, providing predicted image segmentation maps generated by the generator neural network to the critic neural network, training the critic neural network to determine weights for respective pixels of a predicted image segmentation map generated by the generator neural network, wherein the weight for each pixel is used to weight a pixel-wise cross entropy term in an objective function for the critic neural network, and using the weights determined by the critic neural network to train the generator neural network.

Abstract: Systems and methods for image generation are described. Embodiments of the present disclosure receive a text phrase that describes a target image to be generated; generate text features based on the text phrase; retrieve a search image based on the text phrase; and generate the target image using an image generation network based on the text features and the search image.

Abstract: Disclosed are systems, methods, and non-transitory computer-readable media for using adversarial learning for fine-grained image search. An image search system receives a search query that includes an input image depicting an object. The search system generates, using a generator, a vector representation of the object in a normalized view. The generator was trained based on a set of reference images of known objects in multiple views, and feedback data received from an evaluator that indicates performance of the generator at generating vector representations of the known objects in the normalized view. The evaluator including a discriminator sub-module, a normalizer sub-module, and a semantic embedding sub-module that generate the feedback data. The image search system identifies, based on the vector representation of the object, a set of other images depicting the object, and returns at least one of the other images in response to the search query.

Abstract: Embodiments described herein provide for training a machine learning model for automatic organ segmentation. A processor executes a machine learning model using an image to output at least one predicted organ label for a plurality of pixels of the image. Upon transmitting the at least one predicted organ label to a correction computing device, the processor receives one or more image fragments identifying corrections to the at least one predicted organ label. Upon transmitting the one or more image fragments and the image to a plurality of reviewer computing devices, the processor receives a plurality of inputs indicating whether the one or more image fragments are correct. When a number of inputs indicating an image fragment of the image fragments is correct exceeds a threshold, the processor aggregates the image fragment into a training data set. The processor trains the machine learning model with the training data set.

Abstract: Embodiments described herein provide methods and systems for open vocabulary object detection of images. given a pre-trained vision-language model and an image-caption pair, an activation map may be computed in the image that corresponds to an object of interest mentioned in the caption. The activation map is then converted into a pseudo bounding-box label for the corresponding object category. The open vocabulary detector is then directly supervised by these pseudo box-labels, which enables training object detectors with no human-provided bounding-box annotations.

Abstract: Embodiments of the present disclosure provide a quantitative method and system for attention based on a line-of-sight estimation neural network, which improves the stability and training efficiency of the line-of-sight estimation neural network. A few-sample learning method is applied to training of the line-of-sight estimation neural network, which improves generalization performance of the line-of-sight estimation neural network. A nonlinear division method for small intervals of angles of the line of sight is provided, which reduces an estimation error of the line-of-sight estimation neural network. Eye opening and closing detection is added to avoid the line-of-sight estimation error caused by an eye closing state. A method for solving a landing point of the line of sight is provided, which has high environmental adaptability and can be quickly used in actual deployment.

Abstract: An information processing system includes a processor configured to perform: (a) identifying, by inputting photographic image data representing a photographic image of an object into a machine learning model, identification information of the object; and (b) determining, based on the identification information identified in (a), inspection settings used for inspecting the object.

Abstract: Provided are an image sensor module, an image processing system, and an image compression method. The image compression method of compressing image data generated by an image sensor includes: receiving pixel values of a target pixel group of image data on which compression is to be performed, and reference values of reference pixels to be used in compression of the target pixel group; determining an averaging direction in which an averaging calculation is to be performed on target pixel values; averaging the pixel values of target pixels in the averaging direction; generating balance information including compensation values to be applied to the average values based on the reference pixels; and generating a bitstream based on the average values, the balance information, and compression information.

Abstract: Systems, methods, and computer-readable media are provided for efficient control and data utilization between processing components of a system. An method can include obtaining image data captured by an image sensor; prior to a first computing component performing a first set of operations on the image data and a second computing component performing a second set of operations on the image data, determining one or more common operations included in the first set of operations and the second set of operations, wherein the first set of operations is different than the second set of operations; performing the one or more common operations on the image data; and generating an output of the one or more operations for use by the first computing component to perform the first set of operations and the second computing component to perform the second set of operations.

Abstract: An operation method of a server for identifying disaster affected areas. The operation method of the server may include acquiring at least one first disaster image; deriving an affected area from each of the at least one first disaster image and acquiring affected area related information through labeling based on the derived affected area; and training a first learning model using the at least one first disaster image and the affected area related information.

Abstract: A collision warning system and method are provided. The system includes an image capturing apparatus, a sensing apparatus, and a computing apparatus. The computing apparatus analyzes a vehicle side image to generate a plurality of objects and an object coordinate corresponding to each of the objects. The computing apparatus calculates an aerial view coordinate corresponding to each of the objects based on the object coordinates and a pitch angle corresponding to the image capturing apparatus. The computing apparatus calculates a predicted collision time between the vehicle and each of the objects based on a vehicle speed, a turning angle, and the aerial view coordinates. The computing apparatus generates a plurality of warning regions and determines a warning scope corresponding to each of the warning regions based on the predicted collision times.

Abstract: The present invention relates to a neural network for object segmentation. A method for object segmentation using a trained neural network according to an embodiment of the present invention includes receiving a segmentation target image, splitting the target image into unit images having a predetermined size, outputting a unit activation map for the split unit image as a first segmentation result by using the neural network, merging the unit activation map, and outputting a second segmentation result according to the merged unit activation map, in which the neural network is trained by mutually using the entire activation map for the target image and the merged activation map. According to the present invention, it is possible to generate a more accurate object segmentation result, and generate a label using the segmentation result and use the generated label for training of a neural network.

Abstract: The invention describes a method for automatically localizing organ segments in a three-dimensional image comprising the following steps: providing a three-dimensional image showing at least one organ and at least one tubular network comprising a plurality of tubular structures, the organ comprising organ segments; performing automatic separation of the organ from other parts of the image; performing automatic tracing of the tubular network to obtain a branch map; performing automatic analysis of the branch map to identify specific tubular structures; performing automatically assigning regions of the organ to the specific tubular structures to segment the organ into localized organ segments; and outputting the localized organ segments and the traced and analyzed tubular network as image data. The invention further describes a localization arrangement and a medical imaging system.