Abstract: Various implementations disclosed herein include devices, systems, and methods that detect planar regions using depth points from multiple depth sensors (e.g., a sparse array of depth sensors). In some implementations, depth points are obtained from depth sensors, the depth points corresponding to distances of portions of a physical environment from the depth sensors. In some implementations, a subset of the depth points corresponding to a planar region are identified based on positions of the depth sensors corresponding to the depth points of the subset. In some implementations, the planar region is expanded to include an additional depth point or an additional planar region based on expansion criteria, wherein the expansion criteria require that the additional depth point or additional planar region be within a predetermined distance of the planar region.

Abstract: Described is a technique for optimization an image for facial detection. More specifically, described is a process of predicting the location of a face within an image and adjusting image settings based on at least a portion of the predicted location of the face. An image may be adjusted based on the characteristics of a metering region, which may be selected prior to performing facial detection. For example, the metering region may be a specified shape with dimensions equal to a certain percentage of the input image and placed at a specified location. The result of using such a metering region is that the image adjustments may be based on a portion of the face, and therefore, may be optimized for facial detection.

Abstract: Embodiments include a method for object detection in a Light Detection And Ranging (LiDAR) point cloud, the method comprising: placing, by a navigation system, a plurality of anchor points in a two-dimensional Bird's Eye View (BEV) of spatial points represented in a segmented ground surface representation of objects detected by a LiDAR system; extracting, by the navigation system, one or more features from the two-dimensional BEV of the spatial points; proposing, by the navigation system, one or more regions of the two-dimensional BEV of the spatial points for object detection; and performing, by the navigation system, object detections on anchor points of the plurality of anchor points in the proposed one or more regions of the two-dimensional BEV of the spatial points.

Abstract: Apparatuses, systems, and techniques determine a set of grasp poses that would allow a robot to successfully grasp an object that is proximate to at least one additional object. In at least one embodiment, the set of grasp poses is modified based on a determination that at least one of the grasp poses in the set of grasp poses would interfere with at least one additional object that is proximate to the object.

Abstract: A system and method for a scalable depth sensor. The scalable depth sensor having an emitter, a receiver, and a processor. The emitter is configured to uniformly illuminate a scene within a field-of-view of the emitter. The receiver including a plurality of detectors, each detector configured to capture depth and intensity information corresponding to a subset of the field-of-view. The a processor connected to the detector and configured to selectively sample a subset of the plurality of the detectors in accordance with compressive sensing techniques, and provide an image in accordance with an output from the subset of the plurality of the detectors, the image providing a depth and intensity image corresponding to the field-of-view of the emitter.

Abstract: A method for detecting a parallax problem in sensor data from two sensors, wherein the sensors are spaced apart at different positions and at least partly capture the same environment, and one of the sensors provides distance information items. The method includes receiving the acquired sensor data from the sensors; receiving or estimating sensor visual range information items from the other of the sensors; assigning measured values in the acquired sensor data of the one sensor to measured values corresponding to each of them from the other sensor, the assignment taking into account respective imaging conditions; comparing the distance information items and the received or estimated sensor visual range information items on each of the measured values of the sensor data, wherein a parallax problem is detected in response to a distance exceeding a threshold criterion; and outputting a comparison result.

Abstract: An image classification system for tracking a net sport is illustrated. The system includes a visual device configured to capture video footage of a match. The computing device is communicatively connected to the visual device and is configured to receive the video footage of the match from the visual device, translate video footage of the match from the visual device into a data representation by logging an event with a timestamp, and display data representation to a user.

Abstract: A method of fabricating a semiconductor device includes generating a mask based on second layout data obtained by applying an OPC model to first layout data and performing a semiconductor process using the mask on a substrate, obtaining a plurality of pattern images by selecting a plurality of sample patterns from the substrate, selecting sample images corresponding to the sample patterns from each of the first layout data, the second layout data, and simulation data obtained by performing a simulation based on the second layout data, generating a plurality of input images corresponding to the sample patterns by blending the sample images corresponding to the sample patterns, respectively, and generating an error prediction model for the OPC model by training a machine learning model using a data set including the input images and the pattern images.

Abstract: In some implementations, a device may detect edges in an image, and may identify, based on the edges, a rectangle that bounds a document in the image. The device may detect lines in the image, and may identify edge candidate lines by discarding one or more of the lines. The device may identify intersection points where lines, included in the edge candidate lines, intersect with one another. The device may identify corner candidate points by discarding one or more points included in the intersection points, and may identify a corner point included in the corner candidate points. The corner point may be a point, included in the corner candidate points, that is closest to one corner of the bounding rectangle. The device may perform perspective correction on the image of the document based on identifying the corner point.

Abstract: Systems and methods are described for natural language processing of a text sequence. The system can identify a set of text and location information for the set of text in an image. The set of text may correspond to an input sequence space. The system can project embeddings of the text into a latent space for processing. Further, the system can reproject the processed embeddings from the latent space to the input sequence space. The system may perform multiple stages of projecting the embeddings to the latent space and reprojecting the processed embeddings from the latent space to the input sequence space. The system can route the reprojected embeddings to a neural network that can identify class predictions for elements of the set of text.

Abstract: The present application discloses a high-precision mapping method and device, which relates to the field of autonomous driving. A specific implementation includes: acquiring global initial poses of multiple point clouds, where the point clouds are point clouds of a location for which a map is to be built and are collected by a lidar using a multi-circle collection mode; dividing the multiple point clouds into multiple spatial submap graphs according to a spatial distribution relationship of the multiple point clouds; optimizing, for each spatial submap graph, global initial poses of point clouds belonging to the spatial submap graph to acquire global poses of the point clouds in each spatial submap graph; and stitching the multiple spatial submap graphs together according to global poses of the point clouds in the multiple spatial submap graphs to acquire a base graph of the map to be built.

Abstract: The microlens amplitude masks for flying pixel removal in time-of-flight imaging includes systems, devices, methods, and instructions for image depth determination, including receiving an image, adding noise to the image, determining a set of correlation images, each correlation image having a varying phase offset, for each pixel of the image, generating a masked pixel by applying a mask array, and for each masked pixel, determining the depth of the masked pixel to generate a depth map for the image on a per pixel basis.

Abstract: A system and method for generating basic information for positioning and a self-positioning determination device are disclosed. The system includes an information supplying device and a computing device. The information supplying device recognizes an object position and an object category of each of a plurality of reference objects and accordingly generates a reference unique feature value for each reference object, and the computing device generates basic information for positioning according to the reference unique feature values. The self-positioning determination device recognizes an object position and an object category of a current object and accordingly generates a current unique feature value for the current object, and determines a position of itself according to the current unique feature value, the basic information for positioning, and a distance and an angle between the self-positioning determination device and the current object.

Abstract: A processor-implemented method with blur estimation includes: acquiring size information of an input image; resizing the input image to generate a target image of a preset size; estimating a blur of the target image; and estimating a blur of the input image based on the size information of the input image.

Abstract: The present disclosure relates to methods and systems for generating a verified minimum viable range (MVR) of an object. An exposed outer corner and exposed edges of an object may be identified by processing one or more image data. An initial MVR may be generated by identifying opposing parallel edges opposing the exposed edges. The initial MVR may be adjusted, and the adjusted result may be tested to generate a verified MVR.

Abstract: Potentially fraudulent documents can be identified automatically and flagged for further review. Document fraud detection can include a number of fraud detection tests, where if the document fails any of the tests, the document can be flagged for further review. Various tests can be performed that seek to determine whether a document has been altered. In one instance, a financial instrument can be scanned and analyzed to identify one or more layers associated with the financial instrument from a scanned representation, and a check of transaction details can be triggered when the financial instrument is associated with multiple layers.

Abstract: A method of detecting an object depicted in a digital image includes: detecting a plurality of identifying features of the object, wherein the plurality of identifying features are located internally with respect to the object; projecting a location of region(s) of interest of the object based on the plurality of identifying features, where each region of interest depicts content; building and/or selecting an extraction model configured to extract the content based at least in part on: the location of the region(s) of interest, the of identifying feature(s), or both; and extracting the some or all of the content from the digital image using the extraction model. Corresponding system and computer program product embodiments are disclosed. The inventive concepts enable reliable extraction of data from digital images where portions of an object are obscured/missing, and/or depicted on a complex background.

Abstract: A vehicle information detection method, an electronic device and a storage medium are provided, and relates to the technical field of artificial intelligence, in particular to the technical field of computer vision and deep learning. The method includes: determining a bird's-eye view of a target vehicle based on an image of the target vehicle; performing feature extraction on the image of the target vehicle and the bird's-eye view respectively, to obtain first feature information corresponding to the image of the target vehicle and second feature information corresponding to the bird's-eye view of the target vehicle; and determining three-dimensional information of the target vehicle based on the first feature information and the second feature information. According to embodiments of the disclosure, accurate detection of vehicle information can be realized based on a monocular image.

Abstract: The present disclosure provides a logo picture processing method, apparatus, device and medium, and relates to technical field of image processing, and specifically to the technical field of artificial intelligence such as deep learning and computer vision. The logo picture processing method includes: obtaining a logo picture including: a current logo graph and current text information; performing text recognition on the logo picture to obtain the current text information; searching for a picture that matches both the current logo graph and the current text information, to obtain a matched picture. The present disclosure may improve the accuracy of the matched picture of the logo picture and thereby improve the logo picture recognition accuracy.

Abstract: A time-of-flight camera for generating a depth map indicating distance(s) to target(s) includes a processor and multi-pixel time-of-flight image sensor. The camera: (a) determines, at each pixel, a plurality of phase-correlation values associated with at least one exposure duration and at least one phase offset; (b) determines for each pixel an accumulated correlation in response to the plurality of phase-correlation values; and (c) generates the depth map in response to a plurality of the accumulated correlations. A set of accumulated correlations may be determined in response to a plurality of sets of phase-correlation values such that each accumulated correlation is associated with one unique phase offset in response to each set of phase-correlation values being associated with one exposure duration, the depth map being generated in response to a plurality of sets of accumulated correlations. A computer-implemented method of generating a depth map using a time-of-flight image sensor is provided.