Abstract: A method includes receiving, by a computing device, training data to train a neural network, wherein the training data comprises a plurality of inputs and a plurality of corresponding labels. The method also includes mapping, by a representation learner of the neural network, the plurality of inputs to a plurality of feature vectors. The method additionally includes training a kernelized classification layer of the neural network to perform nonlinear classification of an input feature vector into one of a plurality of classes, wherein the kernelized classification layer is based on a kernel which enables the nonlinear classification, and wherein the kernel is selected from a space of positive definite kernels based on application of a nonlinear softmax loss function to the plurality of feature vectors and the plurality of corresponding labels. The method further includes outputting a trained neural network comprising the representation learner and the trained kernelized classification layer.

Abstract: A method calibrates one or more cameras, wherein the one or more cameras acquire images of a scene for different viewpoints, and wherein the images are non-overlapping, by first constructing a 3D model of the scene using a calibration camera that is independent of the one or more cameras, and a simultaneous localization and mapping (SLAM) procedure. 2D-to-3D correspondences between each of the images and the 3D model are determined. Then, calibration parameters of each of the one or more cameras are estimated using a 2D-to-3D registration procedure.

Abstract: A system for reconstructing a three-dimensional (3D) model of a scene including a point cloud having points identified by 3D coordinates includes at least one sensor to acquire a set of images of the scene from different poses defining viewpoints of the images and a memory to store the set of images and the 3D model of the scene. The system also includes a processor operatively connected to the memory and coupled with stored instructions to transform the images from the set of images to produce a set of virtual images of the scene viewed from virtual viewpoints; compare at least some features from the images and the virtual images to determine the viewpoint of each image in the set of images; and update 3D coordinates of at least one point in the model of the scene to match coordinates of intersections of ray back-projections from pixels of at least two images corresponding to the point according to the viewpoints of the two images.

Abstract: Systems and methods for an imaging system having a memory with a historical localization dictionary database having geo-located driving image sequences, such that each reference image is applied to a threshold to produce a binary representation. A sensor to acquire a sequence of input images of a dynamic scene. An encoder to determine, for each input image in the sequence, a histogram of each input image indicating a number of vertical edges at each bin of the input image and to threshold the histogram to produce a binary representation of the input image. A visual odometer to compare the binary representations of each input image and each reference image, by matching an input image against a reference image. Wherein the visual odometer determines a location of the input image based on a match between the input image and the reference image.

Abstract: A image processing system for multi-label semantic edge detection in an image includes an image interface to receive an image of a scene including at least one object, a memory to store a neural network trained for performing a multi-label edge classification of input images assigning each pixel of edges of objects in the input images into one or multiple semantic classes, a processor to transform the image into a multi-label edge-map using the neural network detecting an edge of the object in the image and assigning multiple semantic labels to at least some pixels forming the edge, and an output interface to render the multi-label edge-map.

Abstract: A system for assigning a semantic label to a three-dimensional (3D) point of a point cloud includes an interface to transmit and receive data via a network, a processor connected to the interface, a memory storing program modules including a communicator and a labeler executable by the processor. In this case, the communicator causes the processor to perform operations including establishing, using the interface, a communication with an external map server storing a map defined in a map coordinate system, determining a location on the map corresponding to the 3D point, and querying the map using the location to obtain an information relevant to the semantic label at the location in the map, wherein the labeler causes the processor to determine the semantic label of the 3D point using the information.

Abstract: A method and system estimates a three-dimensional (3D) pose of a sensor by first acquiring scene data of a 3D scene by the sensor. Two-dimensional (2D) lines are detected in the scene data and the 2D lines are matched to 3D lines of a 3D model of the scene to produce matching lines. Then, the 3D pose of the sensor is estimated using the matching lines.

Abstract: Systems and methods for an imaging system having a memory with a historical localization dictionary database having geo-located driving image sequences, such that each reference image is applied to a threshold to produce a binary representation. A sensor to acquire a sequence of input images of a dynamic scene. An encoder to determine, for each input image in the sequence, a histogram of each input image indicating a number of vertical edges at each bin of the input image and to threshold the histogram to produce a binary representation of the input image. A visual odometer to compare the binary representations of each input image and each reference image, by matching an input image against a reference image. Wherein the visual odometer determines a location of the input image based on a match between the input image and the reference image.

Abstract: A image processing system for multi-label semantic edge detection in an image includes an image interface to receive an image of a scene including at least one object, a memory to store a neural network trained for performing a multi-label edge classification of input images assigning each pixel of edges of objects in the input images into one or multiple semantic classes, a processor to transform the image into a multi-label edge-map using the neural network detecting an edge of the object in the image and assigning multiple semantic labels to at least some pixels forming the edge, and an output interface to render the multi-label edge-map.

Abstract: Systems and Methods for controlling a movement of cars of an elevator system. A processor determines for each car an individual waiting time of each hall call. Determines for each pair of hall calls assigned for each car, a pairwise delay over the individual waiting time of each hall call in the pair caused by a joint assignment of the car to the pair of the hall calls. Approximate a cumulative waiting time of an assignment of the cars to the hall calls as a sum of individual waiting times for each hall call with the assigned car and a sum of all pairwise delays determined between all pairs of hall calls assigned to the same car. Determine the assignment of the cars using a greedy optimization algorithm that greedily assigns hall calls to the cars to minimize the approximated cumulative waiting time, and control the movement of the cars.

Abstract: Systems and Methods for controlling a movement of cars of an elevator system. A processor determines for each car an individual waiting time of each hall call. Determines for each pair of hall calls assigned for each car, a pairwise delay over the individual waiting time of each hall call in the pair caused by a joint assignment of the car to the pair of the hall calls. Approximate a cumulative waiting time of an assignment of the cars to the hall calls as a sum of individual waiting times for each hall call with the assigned car and a sum of all pairwise delays determined between all pairs of hall calls assigned to the same car. Determine the assignment of the cars using a greedy optimization algorithm that greedily assigns hall calls to the cars to minimize the approximated cumulative waiting time, and control the movement of the cars.

Abstract: A system for reconstructing a three-dimensional (3D) model of a scene including a point cloud having points identified by 3D coordinates includes at least one sensor to acquire a set of images of the scene from different poses defining viewpoints of the images and a memory to store the set of images and the 3D model of the scene. The system also includes a processor operatively connected to the memory and coupled with stored instructions to transform the images from the set of images to produce a set of virtual images of the scene viewed from virtual viewpoints; compare at least some features from the images and the virtual images to determine the viewpoint of each image in the set of images; and update 3D coordinates of at least one point in the model of the scene to match coordinates of intersections of ray back-projections from pixels of at least two images corresponding to the point according to the viewpoints of the two images.

Abstract: An image registrations includes determining a first binary descriptor of a first key point in a first image, determining a second binary descriptor of a second key point in a second image, determining a weighted Hamming distance between the first binary descriptor and the second binary descriptor, and registering the first key point with the second key point when the weighted Hamming distance is below a noise threshold. At least one element in the first or the second binary descriptor is a result of a comparison of a difference between intensities of at least two pixels of the first or the second image with a threshold. At least two weights of the weighted Hamming distance for comparing at least two elements of the first or the second binary descriptors are different.

Abstract: A method generates a time-series signal indicative of a variation of the environment in vicinity of the vehicle with respect to a motion of the vehicle and determines, using the time-series signal, a trajectory of the vehicle to a location different from a current location of the vehicle, the trajectory of the vehicle is a function of time. The method transmits the trajectory of the vehicle to a remote vehicle and controls a motion of the vehicle according to the trajectory of the vehicle.

Abstract: An image registrations includes determining a first binary descriptor of a first key point in a first image, determining a second binary descriptor of a second key point in a second image, determining a weighted Hamming distance between the first binary descriptor and the second binary descriptor, and registering the first key point with the second key point when the weighted Hamming distance is below a noise threshold. At least one element in the first or the second binary descriptor is a result of a comparison of a difference between intensities of at least two pixels of the first or the second image with a threshold. At least two weights of the weighted Hamming distance for comparing at least two elements of the first or the second binary descriptors are different.

Abstract: A method reconstructs a scene as a three-dimensional (3D) model by first acquiring a sequence of images of the scene with a camera. Then, feature point matches or line matches are extracted from the sequence of images, variables for camera optical centers and 3D points are initialized using random values, and n copies of the variables are made depending on an availability of n constraints. The n copies are projected to satisfy each of the n constraints. Then, the n copies are replaced with averages of the copies and the projecting and the replacing are repeated until convergence to provide the 3D model.

Abstract: A method labels an image of a street view by first extracting, for each pixel, an appearance feature for inferring a semantic label, a depth feature for inferring a depth label. Then, a column-wise labeling procedure is applied to the features to jointly determine the semantic label and the depth label for each pixel using the appearance feature and the depth feature, wherein the column-wise labeling procedure is according to a model of the street view, and wherein each column of pixels in the images includes at most four layers.

Abstract: A method and system for registering data by first acquiring the data from a scene by a sensor at different viewpoints, and extracting, from the data, three-dimensional (3D) points and 3D lines, and descriptors associated with the 3D points and the 3D lines. A first set of primitives represented in a first coordinate system of the sensor is selected, wherein the first set of primitives includes at least three 3D points. A second set of primitives represented in a second coordinate system is selected, wherein the second set of primitives includes any combination of 3D points and 3D lines to obtain at least three primitives. Then, using the first set of primitives and the second set of primitives, the 3D points are registered with each other, and the 3D points with the 3D lines to obtain registered primitives, wherein the registered primitives are used in a simultaneous localization and mapping (SLAM) system.

Abstract: A method reconstructs a scene as a three-dimensional (3D) model by first acquiring a sequence of images of the scene with a camera. Then, feature point matches or line matches are extracted from the sequence of images, variables for camera optical centers and 3D points are initialized using random values, and n copies of the variables are made depending on an availability of n constraints. The n copies are projected to satisfy each of the n constraints. Then, the n copies are replaced with averages of the copies and the projecting and the replacing are repeated until convergence to provide the 3D model.

Abstract: A method and system for registering data by first acquiring the data from a scene by a sensor at different viewpoints, and extracting, from the data, three-dimensional (3D) points and 3D lines, and descriptors associated with the 3D points and the 3D lines. A first set of primitives represented in a first coordinate system of the sensor is selected, wherein the first set of primitives includes at least three 3D points. A second set of primitives represented in a second coordinate system is selected, wherein the second set of primitives includes any combination of 3D points and 3D lines to obtain at least three primitives. Then, using the first set of primitives and the second set of primitives, the 3D points are registered with each other, and the 3D points with the 3D lines to obtain registered primitives, wherein the registered primitives are used in a simultaneous localization and mapping (SLAM) system.