Abstract: A machine vision system using a computer generated virtual reference object is disclosed. When the exact measures of the virtual reference object are known the machine vision system is capable of measuring differences on an absolute scale. In the machine vision system, a virtual reference object is produced based on computer drawings. The computer generated virtual reference object is further processed in order to achieve high photorealistic accuracy. The processing may involve combining portions from image of a manufactured object or computationally producing an image that looks like measured object. When the computer generated virtual reference object is based on the model drawing it does not include inaccuracies caused by manufacturing tolerances.

Abstract: A method for lane and roadway detection uses a multitask DNN architecture including an encoder, a first decoder and a second decoder. The method includes the following steps: providing an input image by an optical detection device, filtering the input image by the encoder, generating a first representation of the lane and/or roadway by the encoder, processing the first representation in the first and second decoders, outputting two different representations from the first and second decoders, combining the two different representations, and outputting identified lanes, lane markings and the roadway.

Abstract: Adapting a machine learning model to process data that differs from training data used to configure the model for a specified objective is described. A domain adaptation system trains the model to process new domain data that differs from a training data domain by using the model to generate a feature representation for the new domain data, which describes different content types included in the new domain data. The domain adaptation system then generates a probability distribution for each discrete region of the new domain data, which describes a likelihood of the region including different content described by the feature representation. The probability distribution is compared to ground truth information for the new domain data to determine a loss function, which is used to refine model parameters. After determining that model outputs achieve a threshold similarity to the ground truth information, the model is output as a domain-agnostic model.

Abstract: A computer-implemented method is disclosed for training one or more machine-learned models. The method can include inputting a first image frame and a second image frame into a feature disentanglement model and receiving, as an output of the machine-learned feature disentanglement model, a state feature and a perspective feature. The method can include inputting the state feature and the perspective feature into a machine-learned decoder model and receiving, as an output of the machine-learned decoder model, the reconstructed image frame. The method can include comparing the reconstructed image frame with a third image frame corresponding with the location and the perspective orientation. The method can include adjusting one or more parameters of the machine-learned feature disentanglement model based on the comparison of the reconstructed image frame and the third image frame.

Abstract: A system may be configured to perform enhanced detection of occluded objects in a multiple object detection system. In some aspects, the system may detect a plurality of current objects in a current video frame, generate initial object identifier mapping information assigning a candidate object identifier to a first object of the plurality of current objects, and detect a multi-object tracking (MOT) inaccuracy condition. Further, the system may generate a current fingerprint corresponding to the first object of the plurality of current objects, identify an assignment error based on comparing the current fingerprint to historic fingerprints associated with a plurality of historic objects detected in a previous frame, generate updated object identifier mapping information based on the assignment error, and display the current video frame based on the updated object identifier mapping information.

Abstract: The present disclosure discloses a method and device for detecting a state of holding a steering wheel, wherein the method includes: a detection result is obtained by detecting a video stream collected from a target vehicle; an associated state of the first ROI information and the second ROI information in the two-dimensional plane carried in the detection result is obtained, wherein the first ROI information is ROI information about the steering wheel of the target vehicle, and the second ROI information is ROI information about the driver's hand of the target vehicle; a current state between the hand of the driver and the steering wheel of the target vehicle is determined according to the associated state. Then, the disclosure solves the technical problem in the related art that it is impossible to detect the grasping state of the steering wheel by the driver while driving.

Abstract: A method for random access for beam failure recovery. In the method, a random access configuration for the beam failure recovery is received. In the event of a beam failure, a random access procedure is performed according to the random access configuration.

Abstract: An image sensor is positioned such that a field-of-view of the image sensor encompasses at least a portion of a rack storing items. The image sensor generates angled-view images of the items stored on the rack. A tracking subsystem determines that a person is within a threshold distance of the rack and receives image frames of the angled-view images. A pixel position of a wrist of the person is determined in at least a subset of the received image frames, thereby determining a set of pixel positions of the wrist. An aggregated wrist position is determined based on the set of pixel positions. If the aggregated wrist position is determined to correspond to a position on a shelf of the rack, a trigger signal is provided indicating a shelf-interaction event has occurred.

Abstract: An image processing method includes determining, based on a coordinate position of a to-be-interpolated sample in a target image, a first coordinate position of the to-be-interpolated sample in a source image, determining m reference samples based on the first coordinate position, determining an interpolation weight of each of the m reference samples for the to-be-interpolated sample based on a spherical distance between a coordinate position of each of the m reference samples and the first coordinate position, and determining a pixel value of the to-be-interpolated sample based on a pixel value corresponding to each of the m reference samples and the interpolation weight of each of the m reference samples for the to-be-interpolated sample to obtain the target image.

Abstract: A system and method for recognition of license plate numbers from Bayer-domain image data is provided. The system acquires Bayer-domain image data of a scene comprising a vehicle and inputs such image data to a Deep Neural Network (DNN) model trained for license plate detection (LPD) and license plate recognition (LPR). The system extracts an LPD result as an output of the DNN model. Based on the LPD result, the system extracts, from the input Bayer-domain image data, a region-of-interest (RoI) comprising a license plate image associated with the vehicle. The system inputs the extracted RoI to the DNN model and extracts, as an output of the DNN model, an LPR result which indicates a confidence of the DNN model in recognition of a license plate number of the vehicle from the input RoI. The system determines the license plate number based on the extracted LPR result.

Abstract: A method includes receiving a query from a user requesting a prediction on future location of a target at a time t>0, receiving a trafficability map wherein the target is located in the trafficability map at an initial time t=0, receiving information about the target including uncertainty on speed of the target, modeling motion of the target using information from the trafficability map including terrain and target mobility on different terrain types to locations on the trafficability map, and answering the query from the user based on output from modeling motion of the target.

Abstract: A method for detecting unmanned aerial vehicles (UAV) includes detecting an unknown flying object in a monitored zone of air space. An image of the detected unknown flying object is captured. The captured image is analyzed to classify the detected unknown flying object. A determination is made, based on the analyzed image, whether the detected unknown flying object comprises a UAV. In response to determining that the detected unknown flying object comprises a UAV, one or more radio signals exchanged between the UAV and a user of the UAV are suppressed until the UAV departs from the monitored zone of air space.

Abstract: A method for operating a vehicle including a vehicle sensing system includes generating a baseline image model of an cabin of the vehicle based on image data of the cabin of the vehicle generated by an imaging device of the vehicle sensing system, the baseline image model generated before a passenger event, and generating an event image model of the cabin of the vehicle based on image data of the cabin of the vehicle generated by the imaging device, the event image model generated after the passenger event. The method further includes identifying image deviations by comparing the event image model to the baseline image model with a controller of the vehicle sensing system, the image deviations corresponding to differences in the cabin of the vehicle from before the passenger event to after the passenger event, and operating the vehicle based on the identified image deviations.

Abstract: A vehicle control device includes a parking space candidate detection unit configured to detect a parking space candidate that is a candidate for a parking space where an own vehicle is parked, using a vehicle peripheral image acquired from a camera; a leaving path detection unit configured to detect, using the vehicle peripheral image, a leaving path on which another vehicle that is parked in an other vehicle parking space located in a periphery of the parking space candidate leaves the other vehicle parking space; and a nuisance parking determination unit configured to perform nuisance parking determination including determining, for each detected parking space candidate, whether the parking space candidate overlaps with the detected leaving path, and thereby classifying the parking space candidate into a nuisance parking space that overlaps with the leaving path or a non-nuisance parking space that avoids overlapping with the leaving path.

Abstract: Methods and apparatus for the analysis of moving objects. In one embodiment, a pitch tracking system is disclosed that includes one or more cameras that have been positioned in a desired location and a computing system containing executable software that is configured to receive imaging data from the one or more cameras that captured the desired location and analyze seam placement and/or finger placement as a function of time for the received imaging data. Computer readable media and methodologies are also disclosed herein.

Abstract: Systems, methods, and processor readable media for distributing digital data and electrical power to a plurality of devices over high-impedance cables are disclosed. Certain embodiments include a gateway device connected to a power source, a first device connected to the gateway device by a cable, the cable being a high-impedance cable having at least two conductive paths, and wherein the first device receives electrical power and digital data from the gateway device via the cable over the same conductive path of the cable, a second device connected to the gateway device by the cable wherein the second device receives power and digital data from the gateway device via the cable over the same conductive path, and wherein the power source provides power to the first and second devices via the cable, and wherein the second device is connected to the gateway device through the first device via a daisy-chain topology.

Abstract: Systems and methods for monitoring posture of a user wearing an exosuit are discussed herein. Exosuits worn by users can monitor several movement factors that characterize the user's movements and posture. The user's posture is identified and analyzed, and feedback is provided to the user based on the analyzed posture.

Abstract: A method for classifying a capture taken by a sensor includes receiving a sequence of a plurality of captures taken by a sensor, each of the captures comprising a plurality of elements; generating, per capture, a plurality of raw probability values, each of the raw probability values being linked to a respective one of a plurality of predetermined classes and indicating the probability that the capture or an element of the capture is associated with the respective class; determining, for a respective one of the captures, a plurality of consolidated probability values in dependence of a plurality of base probability values and a plurality of context probability values, wherein the base probability values represent the raw probability values of the respective capture and the context probability values represent the raw probability values of at least one further capture of the sequence other than the respective capture, the context probability values being normalised according to a normalisation rule; and classi

Abstract: Disclosed herein are an apparatus and method for detecting a broadcast signal. The apparatus includes a bootstrap detection unit for detecting whether a bootstrap signal is included in a received broadcast signal based on a preset first bootstrap window and a machine-learning method, a bootstrap offset estimation unit for searching the broadcast signal for the start point of the bootstrap signal using a preset second bootstrap window, estimating a bootstrap offset based on the machine-learning method, and estimating bootstrap symbols from the bootstrap offset, and a demodulation unit for demodulating information included in the broadcast signal from the bootstrap symbols based on the machine-learning method.

Abstract: Techniques and mechanisms to facilitate a data classification functionality by communicating feedback signals with a spiked neural network. In an embodiment, input signaling, provided to the spiking neural network, results in one or more output spike trains which are indicative of that the input signaling corresponds to a particular data type. Based on the one or more output spike trains, feedback signals are variously communicated each to a respective node of the spiking neural network. The feedback signals variously control signal response characteristics of the nodes. Subsequent output signaling by the spiking neural network, in further response the input signaling, is improved based on the feedback control of nodes' signal responses. In another embodiment, the feedback signals are used to adjust synaptic weight values during training of the spiking neural network.