Abstract: Systems, devices, and methods are provided for depth-guided structure from motion. A system may obtain a plurality of image frames from a digital content item that corresponds to a scene and determine, based at least in part on a correspondence search, a set of 2-D keypoints for the plurality of image frames. A depth estimator may be used to determine a plurality of dense depth map for the plurality of image frames. The set of 2-D keypoints and the plurality of dense depth maps may be used to determine a corresponding set of depth priors. Initialization and/or depth-regularized optimization may be performed using the keypoints and depth priors.

Abstract: Systems, devices, and methods are provided for depth guided structure from motion. A system may obtain a plurality of image frames from a digital content item that corresponds to a scene and determine, based at least in part on a correspondence search, a set of 2-D keypoints for the plurality of image frames. A depth estimator may be used to determine a plurality of dense depth map for the plurality of image frames. The set of 2-D keypoints and the plurality of dense depth maps may be used to determine a corresponding set of depth priors. Initialization and/or depth-regularized optimization may be performed using the keypoints and depth priors.

Abstract: Systems and methods are disclosed for identifying high risk parking lots. High risk parking lots may be, for example, parking lots that pose a higher than average risk of collisions and/or theft. Auto insurance claim data may be analyzed to identify hazardous areas. A virtual navigation of roads within the hazardous area may be identified. Public parking lots within the virtual navigation map may be defined, with each public parking lot determined as either in a hazardous area or not. A vehicle may be determined to be approaching or parking in a parking lot in a hazardous area, and a nearby public parking lot not associated with the hazardous area may be selected instead. A route from a current position to the nearby public parking lot may be generated, and the vehicle may be routed to the nearby public parking lot. As a result, collisions and thefts may be reduced.

Abstract: Systems and methods for control systems for facilitating situational awareness of a vehicle are provided. Sensors located around a vehicle may detect moving objects located in one or more vehicle blind spots. These moving objects pose potentially dangerous situations for the vehicle occupants and for the object. The sensors may further determine that the moving object is a cyclist approaching a vehicle blind spot, and that the vehicle is parked. Processors associated with the vehicle may then direct a corrective action, such as automatically locking the vehicle doors, to prevent a vehicle-cyclist collision. The systems and methods provided herein may thereby make these and similar situations safer.

Abstract: A system and method are provided for improving vehicle awareness and safety by generating and transmitting alerts in response to detecting a hazard in the environment omnidirectional to a vehicle awareness system. Omnidirectional environment data, representing kinematic information pertaining to one or more physically detectable elements omnidirectional to the primary vehicle, is acquired by one or more sensors communicatively coupled to ta vehicle. The system analyzes the omnidirectional environment data to detect if one or more hazards in the omnidirectional environment data, representing a change in the kinetic behavior of the one or more physically detectable elements omnidirectional to the vehicle awareness system, has occurred. When the system detects one or more hazards in the omnidirectional environment data, the system generates and transmits an alert to vehicles, vehicle operators, mobile devices, or pedestrians at risk from the hazard.

Abstract: Systems and methods are disclosed for identifying high risk parking lots. High risk parking lots may be, for example, parking lots that pose a higher than average risk of collisions and/or theft. Auto insurance claim data may be analyzed to identify hazardous areas. A virtual navigation of roads within the hazardous area may be identified. Public parking lots within the virtual navigation map may be defined, with each public parking lot determined as either in a hazardous area or not. A vehicle may be determined to be approaching or parking in a parking lot in a hazardous area, and a nearby public parking lot not associated with the hazardous area may be selected instead. A route from a current position to the nearby public parking lot may be generated, and the vehicle may be routed to the nearby public parking lot. As a result, collisions and thefts may be reduced.

Abstract: Systems and methods are disclosed for educating vehicle drivers. Auto insurance claim data may be analyzed to identify hazardous areas associated with an abnormally high amount or severity of vehicle collisions. A virtual navigation map of roads within the hazardous areas may be built or generated. A common cause of several vehicle collisions at a hazardous area may be identified, and a virtual reconstruction of a scenario involving the common cause and/or a road map of collisions locations of may be created. The virtual reconstruction of the scenario may be displayed on a driver education virtual simulator to enhance driver education and reduce the likelihood of vehicle collisions.

Abstract: An automatic alarm system for a vehicle includes a sensor adapted to detect electromagnetic data, a computer system, and an automatic alarm mechanism. The computer system receives data from the sensor. A computer processor analyzes the data from the sensor to determine whether a pattern from the electromagnetic data has been detected and compares the pattern of electromagnetic data detected by the sensor with known patterns of electromagnetic data emitted by emergency vehicles that are stored in a library. The automatic alarm mechanism activates in response to the processor determining that the comparison matches a known pattern of electromagnetic data for an emergency vehicle resulting in the computer processor sending a first signal to activate an alarm of the vehicle.

Abstract: Systems and methods for providing awareness of emergency vehicles are provided. Sensors located around a vehicle may detect moving objects that pose potentially dangerous situations for the vehicle occupants and for the object. This can cause the doors to lock automatically in order to make the situation safer. Additionally, electromagnetic radiation sensors located around the vehicle may detect the lights and sounds of emergency vehicles. If the electromagnetic data from the electromagnetic radiation sensors match a known pattern of electromagnetic data emitted by emergency vehicles stored in a library an alarm may be sounded through the speakers of the vehicle in order to alert the vehicle occupants of an emergency vehicle in the vicinity.

Abstract: Systems and methods are disclosed for identifying high risk parking lots. High risk parking lots may be, for example, parking lots that pose a higher than average risk of collisions and/or theft. Auto insurance claim data may be analyzed to identify hazardous areas. A virtual navigation of roads within the hazardous area may be identified. Public parking lots within the virtual navigation map may be defined, with each public parking lot determined as either in a hazardous area or not. A vehicle may be determined to be approaching or parking in a parking lot in a hazardous area, and a nearby public parking lot not associated with the hazardous area may be selected instead. A route from a current position to the nearby public parking lot may be generated, and the vehicle may be routed to the nearby public parking lot. As a result, collisions and thefts may be reduced.

Abstract: Systems and methods relate to, inter alia, calculating a number of expected collisions in an area over a time period. The systems and methods may further determine a number of observed collisions in the area over the time period. The systems and methods may further calculate a risk index for the area based upon a comparison between the number of expected collisions and the number of observed collisions. The systems and methods may further select a travel route for a vehicle based upon the risk index.

Abstract: A system and method are provided for improving vehicle awareness and safety by generating and transmitting alerts in response to detecting a hazard in the environment omnidirectional to a vehicle awareness system. Omnidirectional environment data, representing kinematic information pertaining to one or more physically detectable elements omnidirectional to the primary vehicle, is acquired by one or more sensors communicatively coupled to ta vehicle. The system analyzes the omnidirectional environment data to detect if one or more hazards in the omnidirectional environment data, representing a change in the kinetic behavior of the one or more physically detectable elements omnidirectional to the vehicle awareness system, has occurred. When the system detects one or more hazards in the omnidirectional environment data, the system generates and transmits an alert to vehicles, vehicle operators, mobile devices, or pedestrians at risk from the hazard.

Abstract: One or more services may generate audio data and animations of an avatar based on input text. A speech input ingestion (SII) service may identify tags of objects in a virtual environment and associate tags of those objects with words in the input text, which may be stored as metadata in speech markup data. This association may enable an animation service to generate gestures toward objects while animating an avatar, or may be used to create animations or effects of the object. The SII service may analyze input text to identify dialog including multiple speakers associated with the text. The SII service may create metadata to associate certain words with respective speakers (avatars) of those words, which may be processed by the animation service to animate multiple avatars speaking the dialog.

Abstract: An automatic alarm system for a vehicle includes a sensor adapted to detect electromagnetic data, a computer system, and an automatic alarm mechanism. The computer system receives data from the sensor. A computer process analyzes the data from the sensor to determine whether a pattern from the electromagnetic data has been detected, accesses a library of known patterns of electromagnetic data emitted by emergency vehicles, and compares the pattern of electromagnetic data detected by the sensor with the known patterns of electromagnetic data emitted by emergency vehicles that are stored in the library. The automatic alarm mechanism activates in response to the processor determining that the comparison matches a known pattern of electromagnetic data for an emergency vehicle resulting in the computer processor sending a first signal to activate an alarm.

Abstract: Systems and methods for control systems for facilitating situational awareness of a vehicle are provided. Sensors located around a stationary vehicle may detect moving objects that pose potentially dangerous situations for the vehicle occupants and for the object. This can cause the doors to lock automatically in order to make the situation safer. Additionally, electromagnetic radiation sensors located around the vehicle may detect the lights and sounds of emergency vehicles. If the electromagnetic data from the electromagnetic radiation sensors match a known pattern of electromagnetic data emitted by emergency vehicles stored in a library an alarm may be sounded through the speakers of the vehicle in order to alert the vehicle occupants of an emergency vehicle in the vicinity.

Abstract: Systems and methods are disclosed for identifying high risk parking lots. High risk parking lots may be, for example, parking lots that pose a higher than average risk of collisions and/or theft. Auto insurance claim data may be analyzed to identify hazardous areas. A virtual navigation of roads within the hazardous area may be identified. Public parking lots within the virtual navigation map may be defined, with each public parking lot determined as either in a hazardous area or not. A vehicle may be determined to be approaching or parking in a parking lot in a hazardous area, and a nearby public parking lot not associated with the hazardous area may be selected instead. A route from a current position to the nearby public parking lot may be generated, and the vehicle may be routed to the nearby public parking lot. As a result, collisions and thefts may be reduced.

Abstract: Systems and methods for control systems for facilitating situational awareness of a vehicle are provided. Sensors located around a stationary vehicle may detect moving objects that pose potentially dangerous situations for the vehicle occupants and for the object. This can cause the doors to lock automatically in order to make the situation safer. Additionally, electromagnetic radiation sensors located around the vehicle may detect the lights and sounds of emergency vehicles. If the electromagnetic data from the electromagnetic radiation sensors match a known pattern of electromagnetic data emitted by emergency vehicles stored in a library an alarm may be sounded through the speakers of the vehicle in order to alert the vehicle occupants of an emergency vehicle in the vicinity.

Abstract: Techniques for system backup are disclosed. In one embodiment, the techniques may be realized as a method including identifying a particular characteristic of a particular modification of a default automated procedure for selecting files to include in an off-site backup, wherein the identifying is based at least in part on an aggregation of data from many client devices as to how each of many users deviated from the default automated selection procedure; receiving a first system profile for a first client device; identifying the particular characteristic in the first system profile; and based on identifying the particular characteristic in the first system profile, applying a modified default automated selection procedure to the first client device, the modified default automated selection procedure including the particular modification identified with the particular characteristic.

Abstract: A system and method are provided for improving vehicle awareness and safety by generating and transmitting alerts in response to detecting a hazard in the environment omnidirectional to a vehicle awareness system. Omnidirectional environment data, representing kinematic information pertaining to one or more physically detectable elements omnidirectional to the primary vehicle, is acquired by one or more sensors communicatively coupled to ta vehicle. The system analyzes the omnidirectional environment data to detect if one or more hazards in the omnidirectional environment data, representing a change in the kinetic behavior of the one or more physically detectable elements omnidirectional to the vehicle awareness system, has occurred. When the system detects one or more hazards in the omnidirectional environment data, the system generates and transmits an alert to vehicles, vehicle operators, mobile devices, or pedestrians at risk from the hazard.

Abstract: A powder aerosolization apparatus comprises a housing comprising an outlet adapted to be inserted into a user's mouth and one or more bypass air openings. A receptacle support within the housing supports a receptacle containing a powder pharmaceutical formulation. A puncturing mechanism within the housing creates in the receptacle one or more inlet openings and one or more powder outflow openings, wherein the powder outflow openings have a total area of from 0.2 mm2 to 4.0 mm2. Upon a user's inhalation through the outlet, air flows through the one or more bypass air openings and through the receptacle to aerosolize the powder pharmaceutical formulation in the receptacle. In one version, the relative flow parameters between the flow through the one or more bypass openings and the one or more powder outflow openings are selected so that flow of aerosolized pharmaceutical formulation does not occur until a predetermined inhalation flow rate is achieved.