Abstract: A system and method for lean ortho correction for computer models of structures. The system displays an image of a structure on a user interface and projects a structure model onto the image. Next, the system identifies a first world three-dimensional (“3D”) point in the image, a second world 3D point in the image, and a third world 3D point in the image. The system then transforms coordinates of the structure model using the first world 3D point, the second world 3D point, the third world 3D point, and a lean ortho correction algorithm.

Abstract: Machine learning systems and methods for interactive concept searching using attention scoring are provided. The system receives textual data. The system identifies one or more word representations of the textual data. The system further receives a concept. The system determines a score indicative of a likelihood of each of the one or more word representations being representative of the concept using an attention scoring process having a temperature variable. The system generates a dataset for training and evaluation based at least in part on the score. The dataset includes the one or more word representations and concept. The system further processes the dataset to train one or more deep active learning models capable of the interactive concept search.

Abstract: Systems and methods for lossless compression of tabular numeric data are provided. The system can include one or more data compression servers executing data compression system code to compress the tabular numeric data, a storage database to store the compressed tabular numeric data, and one or more data decompression servers to decompress the tabular numeric data for use. The one or more data compression servers, the storage database, and the one or more data decompression servers can communicate via a communication network. The system can receive the uncompressed tabular numeric data from one or more data generation systems, processes the uncompressed tabular numeric data with the data compression system code, and generate a compressed table of numeric information, which can be stored in the database, or later decompressed by the one or more data decompression servers.

Abstract: Computer vision systems and methods for generating building models using three-dimensional sensing and augmented reality (AR) techniques are provided. Image frames including images of a structure to be modeled are captured by a camera of a mobile device such as a smart phone, as well as three-dimensional data corresponding to the image frames. An object of interest, such as a structural feature of the building, is detected using both the image frames and the three-dimensional data. An AR icon is determined based upon the type of object detected, and is displayed on the mobile device superimposed on the image frames. The user can manipulate the AR icon to better fit or match the object of interest in the image frames, and can capture the object of interest using a capture icon displayed on the display of the mobile device.

Abstract: Systems and methods for dynamic computer modeling using incomplete data are provided. The system can yield a total score indicative of an accuracy and reliability of a model for a given application based on incomplete data. The system can receive one or more sets of datasets, classify each dataset among the set of datasets based on a classification component, and determine a normalized score for each dataset based on a value or values of each dataset. If a classification component does not comprise more than one dataset, then the system determines a classification component score for each dataset as the normalized score. If a classification component comprises more than one dataset, the system assigns a weighted data value to each dataset of the classification component. A classification component score is determined for each weighed dataset by applying the weighted data value to the normalized score for each weighed dataset.

Abstract: Systems and methods for aligning digital image datasets to a computer model of a structure. The system receives a plurality of reference images from an input image dataset and identifies common ground control points (“GCPs”) in the reference images. The system then calculates virtual three-dimensional (“3D”) coordinates of the measured GCPs. Next, the system calculates and projects two-dimensional (“2D”) image coordinates of the virtual 3D coordinates into all of the images. Finally, using the projected 2D image coordinates, the system performs spatial resection of all of the images in order to rapidly align all of the images.

Abstract: Systems and methods for computerized loss scenario modeling and data analytics are provided herein. The systems and methods provide customized claim loss analytics to identify questionable insurance claims that require further investigation and include a database for storing claims data related to one or more loss events, a front-end processor configured to generate a user interface with elements configured to allow a user to select one or more business rules and one or more data points for a custom loss scenario rule, and a primary claims analytics processor in communication with the front-end processor that implements the custom loss scenario rule into a production data flow process, thereby impacting the results of data analytics performed by the primary processor.

Abstract: Computer vision systems and methods for localizing image forgery are provided. The system generates a constrained convolution via a plurality of learned rich filters. The system trains a convolutional neural network with the constrained convolution and a plurality of images of a dataset to learn a low level representation of each image among the plurality of images. The low level representation is indicative of a statistical signature of at least one source camera model of each image. The system can determine a splicing manipulation localization by the trained convolutional neural network.

Abstract: Computer vision systems and methods for determining structure features from point cloud data using neural networks are provided. The system obtains point cloud data of a structure or a property parcel having a structure present therein from a database. The system can preprocess the obtained point cloud data to generate another point cloud or 3D representation derived from the point cloud data by spatial cropping and/or transformation, down sampling, up sampling, and filtering. The system can also preprocess point features to generate and/or obtain any new features thereof. Then, the system extracts a structure and/or feature of the structure from the point cloud data utilizing one or more neural networks. The system determines at least one attribute of the extracted structure and/or feature of the structure utilizing the one or more neural networks.

Abstract: A system for modeling a three-dimensional structure utilizing two-dimensional segments comprising a memory and a processor in communication with the memory. The processor extracts a plurality of two-dimensional segments corresponding to the three-dimensional structure from a plurality of images indicative of different views of the three-dimensional structure. The processor determines a plurality of three-dimensional candidate segments based on the extracted plurality of two-dimensional segments and adds the plurality of three-dimensional candidate segments to a three-dimensional segment cloud. The processor transforms the three-dimensional segment cloud into a wireframe indicative of the three-dimensional structure by performing a wireframe extraction process on the three-dimensional segment cloud.

Abstract: A system and method for generating a parametric model of a roof structure comprising a processor in communication with a memory. The system receives a plurality of parameters of each roof component composing the roof structure and performs a geometry creation based on the received plurality of parameters. The system generates a constrained three-dimensional geometry based on an output of the geometry creation, and displays a three-dimensional model of the roof structure based on the constrained three-dimensional geometry.

Abstract: Systems and methods for machine learning of medical records are provided. The system can execute multiple machine learning models on the medical records in parallel using multi-threaded approach wherein each machine learning model executes using its own, dedicated computational thread in order to significantly speed up the time with which relevant information can be identified from documents by the system. The multi-threaded machine learning models can include, but are not limited to, sentence classification models, comorbidity models, ICD models, body parts models, prescription models, and provider name models. The system can also utilize combined convolutional neural networks and long short-term models (CNN+LSTMs) as well as ensemble machine learning models to categorize sentences in medical records. The system can also extract service provider, medical specializations, and dates of service information from medical records.

Abstract: Computer vision systems and methods for determining roof shapes from imagery using segmentation networks are provided. The system obtains an image of a structure from an image database, and determines a flat roof structure ratio and a sloped roof structure ratio of the roof structure using a neural network. Based on segmentation processing by the neural network, the system determines a flat roof structure ratio and a sloped roof structure ratio based on a portion of the roof structure classified as being flat and a portion of the roof structure classified as being sloped. Then, the system determines a ratio of each shape type of the roof structure using a neural network. The system generates a roof structure shape report indicative of a predominant shape of the roof structure and ratios of each shape type of the roof structure.

Abstract: Computer vision systems and methods for generating building models using three-dimensional sensing and augmented reality (AR) techniques are provided. Image frames including images of a structure to be modeled are captured by a camera of a mobile device such as a smart phone, as well as three-dimensional data corresponding to the image frames. An object of interest, such as a structural feature of the building, is detected using both the image frames and the three-dimensional data. An AR icon is determined based upon the type of object detected, and is displayed on the mobile device superimposed on the image frames. The user can manipulate the AR icon to better fit or match the object of interest in the image frames, and can capture the object of interest using a capture icon displayed on the display of the mobile device.

Abstract: A method of evaluating the driving behavior in a vehicle. The method includes determining values of a plurality of parameters of the operation of a first vehicle in a first road segment, determining values of the plurality of parameters for one or more second vehicles in a second road segment having similar properties to those of the first road segment, comparing the determined values of the first vehicle and the one or more second vehicles and providing an evaluation of the driving behavior of the first vehicle, responsive to the comparison.

Abstract: A system and method for lean ortho correction for computer models of structures. The system displays an image of a structure on a user interface and projects a structure model onto the image. Next, the system identifies a first world three-dimensional (“3D”) point in the image, a second world 3D point in the image, and a third world 3D point in the image. The system then transforms coordinates of the structure model using the first world 3D point, the second world 3D point, the third world 3D point, and a lean ortho correction algorithm.

Abstract: Systems and methods for dynamic computer modeling using incomplete data are provided. The system can yield a total score indicative of an accuracy and reliability of a model for a given application based on incomplete data. The system can receive one or more sets of datasets, classify each dataset among the set of datasets based on a classification component, and determine a normalized score for each dataset based on a value or values of each dataset. If a classification component does not comprise more than one dataset, then the system determines a classification component score for each dataset as the normalized score. If a classification component comprises more than one dataset, the system assigns a weighted data value to each dataset of the classification component. A classification component score is determined for each weighed dataset by applying the weighted data value to the normalized score for each weighed dataset.

Abstract: Computer vision systems and methods for supplying missing point data in point clouds derived from stereoscopic image pairs are provided. The system retrieves the at least one stereoscopic image pair from the memory based on a received geospatial region of interest, and processes the at least one stereoscopic image pair to generate a disparity map from the at least one stereoscopic image pair. The system then processes the disparity map to generate a depth map from the disparity map. The depth map is then processed to generate a point cloud from the depth map, such that the point cloud lacks any missing point data. Finally, the point cloud is stored for future use.

Abstract: Computer vision systems and methods for detecting power line hazards from imagery are provided. The system obtains at least one image from an image database having an object and/or a structure present therein, and generates a digital surface model (DSM), a vegetation mask, and a three-dimensional (3D) power line model based on the obtained image. The system generates a digital line model based on the 3D power line and a risk distance indicative of a distance between the 3D power line and vegetation proximate to the 3D power line, and also generates a digital tree model.

Abstract: Systems and methods for aligning digital image datasets to a computer model of a structure. The system receives a plurality of reference images from an input image dataset and identifies common ground control points (“GCPs”) in the reference images. The system then calculates virtual three-dimensional (“3D”) coordinates of the measured GCPs. Next, the system calculates and projects two-dimensional (“2D”) image coordinates of the virtual 3D coordinates into all of the images. Finally, using the projected 2D image coordinates, the system performs spatial resection of all of the images in order to rapidly align all of the images.