According to one embodiment, an antenna is configured to irradiate a first electromagnetic wave of a wavelength of 1 mm to 30 mm to a first position in an area in which at least one of a target person or a belonging of the target person is present, and irradiate a second electromagnetic wave of a wavelength of 1 mm to 30 mm to a second position in the area. Processor circuitry is configured to obtain a first reflection intensity of the first electromagnetic wave on the first position, and obtain a second reflection intensity of the second electromagnetic wave on the second position, and determine a degree of danger relating to a possibility that the target person possesses a dangerous article.

Provided is a server of a checking system based on a millimeter wave security inspection device, connected to a security inspection instrument through a switching system, the server includes a memory and a processor, the memory stores instructions, and the instructions, when executed by the processor, cause the processor to: receive a scanned image of an object and an ATR image interpretation result of the object from the security inspection instrument, wherein the ATR image interpretation result is obtained by the security inspection instrument performing an automatic threat recognition (ATR) image interpretation on the scanned image, and perform a manual inspection task on the object and perform an image interpretation task on the scanned image according to a working mode, wherein the working mode indicates whether the checking system is provided with a manual inspection station and an image interpretation station.

An automatic generation system of a training image and a method thereof are provided. The disclosure generates a training image and records the target category and the target position. The disclosure adds the target image to the container image as a candidate image, calculates a reliability of the candidate image, and repeatedly executes the process until the reliability of the candidate image meets a threshold condition for generating the training image. The disclosure is able to generate the training images automatically, and the recognition difficulty of the training image is adjustable by the user, so as to be suitable for customized recognition training.

A method includes receiving, from a plurality of magnetic field receivers including magnetic sensors, data characterizing samples obtained by the plurality of magnetic field receivers, the samples of a combination of a first magnetic field and a second magnetic field resulting from interaction of the first magnetic field and an object; determining, using the received data, a polarizability index of the object, the polarizability index characterizing a magnetic polarizability property of the object; classifying, using the determined polarizability index, the object as threat or non-threat; and providing the classification. Related apparatus, systems, techniques, and articles are also described.

Embodiments described herein relate to a threat detection system that shows a user an incident as it develops in real time by leveraging artificial intelligence (AI) to more accurately focus cameras and highlight the areas of concern within those feeds, providing a much more efficient user interface to the operator. These annotated feeds and feed focused triggering events can also be connected to third party systems. This timeline of events and evidence (small annotated clip of detection when available) is archived and can be reviewed at a later date, containing an accurate timeline of the incident as it progressed.

An intelligent assistant records speech spoken by a first user and determines a self-selection score for the first user. The intelligent assistant sends the self-selection score to another intelligent assistant, and receives a remote-selection score for the first user from the other intelligent assistant. The intelligent assistant compares the self-selection score to the remote-selection score. If the self-selection score is greater than the remote-selection score, the intelligent assistant responds to the first user and blocks subsequent responses to all other users until a disengagement metric of the first user exceeds a blocking threshold. If the self-selection score is less than the remote-selection score, the intelligent assistant does not respond to the first user.

In an embodiment, a method for generating contrastive information for a classifier prediction comprises receiving image data representative of an input image, using a deep learning classifier model to predict a first classification for the input image, evaluating the input image using a plurality of classifier functions corresponding to respective high-level features to identify one or more of the high-level features absent from the input image, and identifying, from among the high-level features absent from the input image, a pertinent-negative feature that, if added to the input image, will result in the deep learning classifier model predicting a second classification for the modified input image, the second classification being different from the first classification. In an embodiment, the method includes creating a pertinent-positive image that is a modified version of the input image that has the first classification and fewer than all superpixels of the input image.

A device may provide a police assistance device to a person. The police assistance device includes hardware for interacting with a person independent of a police officer such as a processor, memory, a display, a camera, a microphone, and a speaker. The devices and methods may connect the suicide attempter with a chosen contact or a mental health provider. The devices and methods may help with traffic stops. The devices and methods may help with weapon detection.

The present invention relates to a system and method for inspecting object using a plurality of interlacing radiation energies. The system comprising a radiation module configured for producing and capturing radiation in multiple energy levels to scan the content of the cargo and converting the captured radiation into a plurality of images; and a controller configured for signalling the radiation module to start or to stop producing radiation and for controlling the energy level and pulse frequency of the radiation produced by the radiation module. The system further comprising a processor configured for determining whether the cargo contains any contraband or not by analysing the plurality of images, classifying the cargo based on types of materials and substance groups and highlighting region on an analysed image of the same substance by bounding perimeter of the object within a material-colour image for the material.

In an embodiment, a method for generating contrastive information for a classifier prediction comprises receiving image data representative of an input image, using a deep learning classifier model to predict a first classification for the input image, evaluating the input image using a plurality of classifier functions corresponding to respective high-level features to identify one or more of the high-level features absent from the input image, and identifying, from among the high-level features absent from the input image, a pertinent-negative feature that, if added to the input image, will result in the deep learning classifier model predicting a second classification for the modified input image, the second classification being different from the first classification. In an embodiment, the method includes creating a pertinent-positive image that is a modified version of the input image that has the first classification and fewer than all superpixels of the input image.