Abstract: The present disclosure relates to training and utilizing an image exposure transformation network to generate a long-exposure image from a single short-exposure image (e.g., still image). In various embodiments, the image exposure transformation network is trained using adversarial learning, long-exposure ground truth images, and a multi-term loss function. In some embodiments, the image exposure transformation network includes an optical flow prediction network and/or an appearance guided attention network. Trained embodiments of the image exposure transformation network generate realistic long-exposure images from single short-exposure images without additional information.

Abstract: A method, an apparatus and an electronic device for processing point cloud data and a computer readable storage medium are disclosed. The method includes: receiving first point cloud data acquired by a laser scanner; classifying the first point cloud data to obtain second point cloud data which is classified; judging if the second point cloud data at least comprises target point cloud data, and whether a distance between other point cloud data in the second point cloud data and the target point cloud data is smaller than a first preset threshold value; if yes, determining the other point cloud data as hazardous point cloud data.

Abstract: Provided are a processing method and system for a convolutional neural network, and a computer-readable medium, the processing method includes training a generator and training a discriminator, wherein training a generator includes: extracting a low-resolution color image from a high-resolution color image; training parameters of a generator network, by using the low-resolution color image and a noise image as an input image, based on parameters of a discriminator network, and reducing a generator cost function; training a discriminator includes: inputting an output image of the trained generator network and the high-resolution color image to the discriminator network, respectively; training parameters of the discriminator network by reducing a discriminator cost function (S204) the generator cost function and the discriminator cost function represent a degree in which the output image of the generator network corresponds to the high-resolution color image.

Abstract: This disclosure relates to techniques for generating robust depth estimations for captured images using semantic segmentation. Semantic segmentation may be defined as a process of creating a mask over an image, wherein pixels are segmented into a predefined set of semantic classes. Such segmentations may be binary (e.g., a ‘person pixel’ or a ‘non-person pixel’) or multi-class (e.g., a pixel may be labelled as: ‘person,’ ‘dog,’ ‘cat,’ etc.). As semantic segmentation techniques grow in accuracy and adoption, it is becoming increasingly important to develop methods of utilizing such segmentations and developing flexible techniques for integrating segmentation information into existing computer vision applications, such as depth and/or disparity estimation, to yield improved results in a wide range of image capture scenarios.

Abstract: There are provided system and method of classifying defects in a specimen. The method includes: obtaining one or more defect clusters detected on a defect map of the specimen, each cluster characterized by a set of cluster attributes comprising spatial attributes including spatial density indicative of density of defects in one or more regions accommodating the cluster, each given defect cluster being detected at least based on the spatial density thereof meeting a criterion. The defect map also comprises non-clustered defects. Defects of interest (DOI) are identified in each cluster by performing respective defect filtrations for each cluster and non-clustered defects.

Abstract: The present invention relates to a method, device, and robot for classifying an outlier during object recognition learning using artificial intelligence. The method or device for classifying an outlier during object recognition learning according to an embodiment of the present invention sets an inlier region and an outlier region through learning using unlabeled data and labeled data.

Abstract: An apparatus includes an interface and a processor. The interface may be configured to receive video frames corresponding to an interior of a vehicle. The processor may be configured to perform video operations on the video frames to detect objects in the video frames, detect a driver based on the objects detected in the video frames, detect a use of an electronic device by the driver and generate a notification signal. The notification signal may be configured to warn the driver about using the electronic device in the vehicle. The notification signal may be generated if the use of the electronic device is unauthorized. The processor may comprise a dedicated hardware module configured to perform feature detection to calculate descriptors for determining a likelihood that pixels of the video frames belong to the objects that correspond to the use of the electronic device by the driver.

Abstract: During operation, an electronic device may capture images using multiple image sensors having different fields of view and positions. Then, the electronic device may determine, based at least in part on an apparent size of an anatomical feature in the images (such as an interpupillary distance) and a predefined or predetermined size of the anatomical feature, absolute motion of at least a portion of the individual along a direction between at least the portion of the individual and the electronic device. Moreover, the electronic device may compute based at least in part on an estimated distance along the direction corresponding to the apparent size and the predefined or predetermined size and angular information associated with one or more objects in the images relative to the positions, absolute motion of at least the portion of the individual in a plane that is perpendicular to the direction.

Abstract: A system and methods for a CODEC driving a real-time light field display for multi-dimensional video streaming, interactive gaming and other light field display applications is provided applying a layered scene decomposition strategy. Multi-dimensional scene data is divided into a plurality of data layers. One or more polygons representative of corresponding portions of objects in the scene are obtained for each layer and used to determine a view-independent representation. This view independent representation and data layers are sampled using a plenoptic sampling scheme and rendered using hybrid rendering to encode light fields corresponding to each data layer. The resulting compressed, (layered) core representation of the multi-dimensional scene data is produced at predictable rates, reconstructed and merged at the light field display in real-time by applying view synthesis protocols, including edge adaptive interpolation, to reconstruct pixel arrays in stages (e.g.

Abstract: Improved watermarking techniques for text content are disclosed. An example methodology implementing the techniques includes selecting a sequence of text characters to form a watermark and representing at least one text character of the sequence of text characters by a code which, when inserted into text content, does not affect the appearance of the text content. The methodology also includes embedding the code which represents the at least one text character of the watermark into text content so that the code enables identification of the at least one text character upon extraction of the code from the text content.

Abstract: A method includes receiving a first depth map that includes a plurality of first pixel depths and a second depth map that includes a plurality of second pixel depths. The first depth map corresponds to a reference depth scale and the second depth map corresponds to a relative depth scale. The method includes aligning the second pixel depths with the first pixel depths. The method includes transforming the aligned region of the second pixel depths such that transformed second edge pixel depths of the aligned region are coextensive with first edge pixel depths surrounding the corresponding region of the first pixel depths. The method includes generating a third depth map. The third depth map includes a first region corresponding to the first pixel depths and a second region corresponding to the transformed and aligned region of the second pixel depths.

Abstract: An optical dimensioning system includes one or more light emitting assemblies configured to project one or more predetermined patterns on an object; an imaging assembly configured to sense light scattered and/or reflected off the object, and to capture an image of the object while the patterns are projected; and a processing assembly configured to analyze the image of the object to determine one or more dimension parameters of the object. The light emitting assembly may include a single piece optical component configured for producing a first pattern and second pattern. The patterns may be distinguishable based on directional filtering, feature detection, feature shift detection, or the like. A method for optical dimensioning includes illuminating an object with at least two detectable patterns; and calculating dimensions of the object by analyzing pattern separate of the elements comprising the projected patterns. One or more pattern generators may produce the patterns.

Abstract: Systems and method for automated driving assistance are described. A driving assistant system may include a first sensor unit configured to provide an input image including another vehicle and a background; an image processor configured to generate a plurality of pyramid images by down-sampling the input image, acquire a depthmap including depth values from the first sensor unit to the other vehicle and the background, generate a plurality of pieces of mask data with different average depths of the depth values based on the depthmap, and output a plurality of masked images representing different average distances from the first sensor unit based on the plurality of pieces of mask data and the plurality of pyramid images; a feature extractor configured to output feature data of each of the plurality of masked images; and a detector configured to detect the other vehicle included in the input image based on the feature data.

Abstract: Methods, computer program products, and systems are presented. The method computer program products, and systems can include, for instance: obtaining from a user one or more data queries; identifying a product of interest in response to the one or more data query; examining a plurality of product records to determine a set of related products that are related to the product of interest, wherein the examining includes performing image analysis to extract one or more product topic classifier from product image data representing one or more product; and providing one or more output in response to the examining.

Abstract: A document inspection system performs texture analysis on a portion of a first image that represents a first document and on a portion of a second image that represents a second document; performing. Feature vectors for the regions from the first and second documents are compared. If the feature vectors have a threshold degree of similarity, they may be identified as being sufficiently related. This may have applicability, for example, in determining with a first document is valid because reflects sufficiently-similar characteristics to those of a reference document.

Abstract: A system and method for reflection removal of an image from a dual-pixel sensor. The image including a left view and a right view. The method including: determining a first gradient of the left view and a second gradient of the right view; determining disparity between the first gradient and the second gradient using a sum of squared differences (SSD); determining a confidence value at each pixel using the SSD; determining a weighted gradient map using the confidence values; minimizing a cost function to estimate the background layer, the cost function including the weighted gradient map, wherein the image includes the background layer added to a reflection layer; and outputting at least one of the background layer and the reflection layer.

Abstract: Feedback data useful in prosthodontic procedures associated with the intra oral cavity is provided. First, a 3D numerical model of the target zone in the intra oral cavity is provided, and this is manipulated so as to extract particular data that may be useful in a particular procedure, for example data relating to the finish line or to the shape and size of a preparation. The relationship between this data and the procedure is then determined, for example the clearance between the preparation and the intended crown. Feedback data, indicative of this relationship, is then generated, for example whether the preparation geometry is adequate for the particular type of prosthesis.

Abstract: In one embodiment, a set of feature vectors can be derived from any biometric data, and then using a deep neural network (“DNN”) on those one-way homomorphic encryptions (i.e., each biometrics' feature vector) can determine matches or execute searches on encrypted data. Each biometrics' feature vector can then be stored and/or used in conjunction with respective classifications, for use in subsequent comparisons without fear of compromising the original biometric data. In various embodiments, the original biometric data is discarded responsive to generating the encrypted values. In another embodiment, the homomorphic encryption enables computations and comparisons on cypher text without decryption. This improves security over conventional approaches. Searching biometrics in the clear on any system, represents a significant security vulnerability. In various examples described herein, only the one-way encrypted biometric data is available on a given device.

Abstract: A vehicle-mounted device is configured to identify another vehicle from an image generated by photographing the surroundings of a host vehicle using a photographing device. The vehicle-mounted device is configured to detect identification information of the another vehicle on the basis of a signal received via inter-vehicle communication with a vehicle-mounted device of the another vehicle located in a communication range of the inter-vehicle communication. The vehicle-mounted device is configured to determine whether the identification information corresponding to the another vehicle identified from an image has been acquired.

Abstract: An electronics system may include a substrate, an electronic device bonded to the substrate, a plurality of photoluminescent particles disposed on the electronic device, an illuminator, a sensor, and a control module. The illuminator can illuminate the electronic device. The sensor can capture a first set of positions of the photoluminescent particles on the electronic device when the electronic device is not operating under a load and a second set of positions of the photoluminescent particles when the electronic device is operating under a load. The control module can determine thermomechanical stress on the electronic device based at least in part on a difference between the first set of positions and the second set of positions.