Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for feature selection for object tracking. One of the methods includes: obtaining first feature points of an object in a first image of a scene captured by a camera; obtaining a second image of the scene captured by the camera after the first image was captured; determining whether a motion prediction of the object is available that indicates an area of the second image where the object is likely located; in response to determining that the motion prediction of the object is available, identifying, in the area of the second image where the object is likely located, second feature points that satisfy a similarity threshold for the first feature points in the first image; and detecting the object in the second image using the identified second feature points.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for weapon object detection and classification and pose aggression analysis. One of the methods includes generating, for an image that depicts a person, a mapped pose model for the person that represents a pose of the person; predicting a first likelihood that the person is holding a weapon using at least some data from the mapped pose model and at least some data from the image; determining, from a plurality of pose types, a pose type for the person using the mapped pose model; determining an action to perform for the person using the first likelihood that the person is holding a weapon and the pose type for the person; and transmitting instructions to cause a device to perform the action.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for detecting motion of an object or event of interest. One of the methods includes maintaining a first image captured by a camera; maintaining, for each of one or more second images, a corresponding difference image generated from the first image and the corresponding second image; providing, to a deep motion detector trained to detect motion for an object of interest and to cause the deep motion detector to generate output, the one or more difference images and color image data for the first image; receiving, from the deep motion detector, output that indicates whether the first image likely depicts motion for an object of interest; and performing one or more automated actions using the output that indicates whether the first image likely depicts motion for an object of interest.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for receiving multiple images from a camera, each image of the multiple images representative of a detection of an object within the image. For each image of the multiple images the methods include: determining a set of detected objects within the image, each object defined by a respective bounding box, and determining, from the set of detected objects within the image and ground truth labels, a false detection of a first object. The methods further include determining that a target object threshold is met based on a number of false detections of the first object in the multiple images, generating, based on the number of false detections for the first object meeting the target object threshold, an adversarial mask for the first object, and providing, to the camera, the adversarial mask.

Abstract: Methods, systems, and apparatus for ground plane filtering of video events are disclosed. A method includes obtaining a first set of images of a scene from a camera; determining a ground plane from the first set of images of the scene; obtaining a second set of images of the scene after the first set of images of the scene is obtained; determining that movement shown by a group of pixels in the second set of images of the scene satisfies motion criteria; determining that the ground plane corresponds with at least a portion of the group of pixels; and in response to determining that movement shown by the group of pixels in the second set of images of the scene satisfies motion criteria, and that the ground plane corresponds with at least a portion of the group of pixels, classifying the group of pixels as showing ground plane based motion.

Abstract: Disclosed are methods, systems, and apparatus for monitoring delivered packages using video. A method includes obtaining a notification of a delivery of a package at a property that includes a first image depicting the package; obtaining a second image captured by a camera at the property; determining, using the first image and the second image, that the package has been delivered to the property; generating a package model that represents an appearance of the package; and tracking a location of the package for use in determining whether to provide an alert to a device about the location of the package. Determining that the package has been delivered to the property comprises: comparing the first image to a model of a scene of the property; determining that the first image satisfies similarity criteria for matching the model of the scene; and determining that the second image likely depicts the package.

Abstract: Methods, systems, and apparatus for camera detection of human activity with co-occurrence are disclosed. A method includes detecting a person in an image captured by a camera; in response to detecting the person in the image, determining optical flow in portions of a first set of images; determining that particular portions of the first set of images satisfy optical flow criteria; in response to determining that the particular portions of the first set of images satisfy optical flow criteria, classifying the particular portions of the first set of images as indicative of human activity; receiving a second set of images captured by the camera after the first set of images; and determining that the second set of images likely shows human activity based on analyzing portions of the second set of images that correspond to the particular portions of the first set of images classified as indicative of human activity.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for fast user enrollment for facial recognition using face clustering. A property monitoring system may reduce user input requirements by automatically performing face clustering, and requesting a user identify face images by cluster. Presenting clusters of faces to a user may allow the user to more quickly indicate multiple face images are of a user, and allow the property monitoring system to prioritize which clusters of faces are to be reviewed by the user. Accordingly, the property monitoring system may reduce processing and requests for user input by clustering faces.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for determining monitoring system actions using contextual information. One of the methods includes providing, to an artificial intelligence model and for an event at a property, contextual information that includes a) first data representing sensor data for the event, b) a role of a person for whom notification instructions are sent, c) an event type for the event, and d) activity data that indicates an activity in which the person is likely involved; in response to providing the contextual information, receiving, from the artificial intelligence model, output that indicates an action for the event; and sending, to a device, instructions to cause the device to perform the action.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for processing event data on a device at a property. One of the methods includes determining, for sensor data maintained by a battery powered device at a property, whether the battery powered device or another device for the property should process the sensor data; and using a result of the determination whether the battery powered device or the other device should process the sensor data, sending an instruction to the battery powered device to cause the battery powered device to process the sensor data or send the sensor data to the other device.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for object detection. One of the methods includes determining, using first sensor data, a detection result on whether to trigger an event alerting a presence of an object in a target area by executing one or more models; determining, using second sensor data, a ground truth for the event that indicates whether an object is present in the target area; determining a difference value by comparing the detection result and the ground truth; adjusting at least one parameter of the one or more models in response to determining that the difference value does not satisfy the one or more threshold criteria; and determining a new detection result on whether to trigger a second event by executing the one or more models with adjusted parameters using new first sensor data.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer-storage media, for visual authentication. In some implementations, a method may include obtaining images of a person at a property; detecting a discontinuity in an appearance of the person in the images of the person at the property; determining that the discontinuity does not correspond to a known occlusion; and providing an indication of a potential spoofing attack.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for object detection. One of the methods includes determining, using first sensor data, a detection result on whether to trigger an event alerting a presence of an object in a target area by executing one or more models; determining, using second sensor data, a ground truth for the event that indicates whether an object is present in the target area; determining a difference value by comparing the detection result and the ground truth; adjusting at least one parameter of the one or more models in response to determining that the difference value does not satisfy the one or more threshold criteria; and determining a new detection result on whether to trigger a second event by executing the one or more models with adjusted parameters using new first sensor data.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for detecting packages delivering in the camera's blindspot. One of the methods includes detecting, using one or more images captured by a camera at a property, movement in an area of interest i) at the property, ii) that is included in a field of view of the camera and iii) was generated using historical data for packages delivered to the property; determining, using the detected movement in the area of interest, that a package was likely delivered; and in response to determining that the package was likely delivered, performing one or more automated actions for the package.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for spatial motion attention for intelligent video analytics. One of the methods includes: obtaining an input image of a region; generating a motion image that characterizes a difference between a value of a pixel at the pixel location in the input image and a value of a pixel at the pixel location in the reference image; generating a feature map using the input image; generating, using the motion image and the feature map, a motion enhanced feature map that has, for one or more pixels that likely indicate movement, a first value that a) indicates that the corresponding pixel in the motion enhanced feature map likely indicates movement and b) is different from a second value for a corresponding pixel in the feature map; and analyzing the motion enhanced feature map.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for controlling access to a premises. One of the methods includes detecting a person within a threshold distance of a door; determining that there is a threshold likelihood that the person may access the door; providing instructions that indicate one or more criteria for access to the door; determining that a threshold quantity of the one or more criteria are satisfied; and in response to determining that the threshold quantity of the one or more criteria are satisfied, providing access to the door.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for receiving multiple images from a camera, each image of the multiple images representative of a detection of an object within the image. For each image of the multiple images the methods include: determining a set of detected objects within the image, each object defined by a respective bounding box, and determining, from the set of detected objects within the image and ground truth labels, a false detection of a first object. The methods further include determining that a target object threshold is met based on a number of false detections of the first object in the multiple images, generating, based on the number of false detections for the first object meeting the target object threshold, an adversarial mask for the first object, and providing, to the camera, the adversarial mask.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer-storage media, for visual authentication. In some implementations, a method may include obtaining images of a person at a property; detecting a discontinuity in an appearance of the person in the images of the person at the property; determining that the discontinuity does not correspond to a known occlusion; and providing an indication of a potential spoofing attack.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for objection classification. One of the methods includes: obtaining a first set of images of objects that have a likelihood of being at a property that satisfies a likelihood threshold; generating, for each object, a binary classifier from a set of images of the respective object; determining, using at least one of the binary classifiers, that an image of an unknown object was classified as an object from the objects; in response to determining, using the binary classifiers, that the image of the unknown object was classified as an object from the objects, selecting a second set of images of unknown objects that does not include the image; and generating a multiclass classifier for use classifying objects using i) the first set as respective classes and ii) the second set that does not include the image.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a storage device, for monitoring a vehicle are disclosed. A vehicle monitoring system for monitoring a vehicle includes one or more processors and one or more computer storage media storing instructions that are operable, when executed by the one or more processors, to cause the one or more processors to perform operations. The operations include: receiving, from a vehicle occupancy monitor and by the vehicle monitoring system, vehicle occupancy data that indicates presence of a vehicle occupant; receiving, by the vehicle monitoring system and from a sensor, sensor data that indicates an attribute of the vehicle; analyzing, by the vehicle monitoring system, the vehicle occupancy data and the sensor data; and based on analyzing the vehicle occupancy data and the sensor data, performing a vehicle monitoring system action.