Abstract: A process for fixed camera and unmanned mobile device collaboration is disclosed in order to improve identification of an object of interest. A first point of view (POV) of a captured object is obtained and it is determined, with a first level of certainty, that the captured first POV of the object matches a stored object of interest. A dispatch instruction and intercept information is then wirelessly broadcast for receipt by camera-equipped unmanned mobile vehicles within a broadcast range for identifying and intercepting the object. Subsequently, a captured second POV of the first captured object is received via the one or more camera-equipped unmanned mobile vehicles. The captured second POV of the captured object is used to determine, with a second level of certainty, that the captured object matches the stored object of interest.

Abstract: Systems for forensic steganography. A server is interfaced with storage facilities that store an object accessible by two or more users, each of which users are associated with respective profiles comprising one or more user-specific attributes. A method detects a user request to view the object. User-specific attributes are encoded into a steganographic message, which is formatted for saving into one or more locations in the object, thus generating a protected object. The protected object is delivered to the requesting user. Encoding, application and formatting techniques are configured to make the steganographic message undetectable by human viewing of the protected object. A web crawler or other policing technique can detect misappropriation in the form of unauthorized dissemination by detecting the presence of the encoded steganographic message embedded in the protected object.

Abstract: The present disclosure relates to a process and a system for assessing the performance of cleaning operations by utilizing an automatic read out routine for test strips subjected to a cleaning operation. Furthermore, the present disclosure relates to a computer program product capable of performing the process.

Abstract: Three-dimensional (3D) depth imaging systems and methods are disclosed for automatic container door status recognition. A 3D-depth camera captures 3D image data of a shipping container located in a predefined search space during a shipping container loading session. A container recognition application (app) receives the 3D image data, and determines therefrom a container point cloud representative of the shipping container, and a 3D image door area data subset defining one or more door areas of one or more corresponding doors of the shipping container. The container recognition app determines, from the 3D image door area data subset, a status type, of each of the one or more corresponding doors, selected from one of: (a) a door opened status type, (b) a door closed status type, or (c) a door partially closed status type.

Abstract: An unmanned insurance drone can include a drone body and a sensor unit disposed on the drone body to collect sensor data. An on-board data processor converts the sensor data into insurance related information, and a wireless communication unit in communication with the data processor to transmit the insurance related information. In another example, the data processor may not be on the drone but remotely located. The location can be with the pilot or a control collection location. If the insurance related information is separate from the drone, than the wireless communication unit can transmit the raw sensor data to the processor.

Abstract: To provide an image sensor capable of guaranteeing consistency between acquired image information and picked-up image information. An image sensor 30 includes an image information processing unit 4 that forms integrated information in which image sensor identification information capable of identifying the image sensor 30 and image information obtained by an analog/digital conversion unit 25 are associated with each other, and an image information output unit 24 that outputs the integrated information to an external unit.

Abstract: Provided is a method of embedding and extracting watermark data that is robust against geometric distortion and low-quality photographing and for which the probability of successfully extracting watermark data for an original image is high, while the probability of successfully extracting the watermark data in the case of unauthorized copying is seriously impaired. The data embedding method according to an embodiment of the present invention comprises a step of converting the noise-based image using watermark data, and a step of adjusting the original image using the converted noise-based image.

Abstract: Systems and methods are provided for providing targeted marketing to goods and services provides and consumer resource management services to consumers. An example system and method for targeted marketing comprises collecting transaction data from point-of-sale (POS) terminals and using a consumer identifier in the transaction data to access stored information about the consumer. This information may be used to target offers and advertisement to the consumer. In an example system for consumer resource management, a consumer may configure a consumer account on the enterprise infrastructure via a web-site. The consumer may use the consumer account to purchase and configure gift cards that may be used for purchasing goods and services. A universal transaction identifier may be associated with the consumer account and used to purchase goods and services from more than one selected goods and services providers.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for locating a client terminal are provided. One of the methods includes: obtaining, by an audio signal playing terminal, identification information of an object associated with the audio signal playing terminal; watermarking, by the audio signal playing terminal, a to-be-played audio signal based on the identification information to obtain an audio signal having a watermark; and playing, by the audio signal playing terminal, the audio signal having the watermark for a client terminal to collect the audio signal having the watermark and send the audio signal having the watermark or the identification information to a server to enable the server to obtain the identification information and determine a position of the object based on a correspondence between the identification information and the position of the object and thereby to determine a position of the client terminal.

Abstract: A method of inserting additional data into a moving image stream, according to some embodiments of the present invention, includes receiving a first coded bitstream, electing a first picture PA into which additional data is to be inserted from the first coded bitstream, generating a first picture PA? by inserting the additional data into the selected first picture PA, selecting a second picture PB directly referring to the first picture PA, generating a second picture PB? by transforming and coding the selected second picture PB in a manner that decoding the second picture PB? obtains the second picture PB with reference to the first picture PA? and generating a data sequence based on the first picture PA? and the second picture PB.

Abstract: In a computer-implemented method, one or more digital aerial images of a property of a current or potential policyholder may be received. The digital aerial image(s) (e.g., camera images, spectrometer images, LiDAR images, etc.) may be processed according to one or more image analysis techniques to determine one or more features of the property. The property feature(s) (e.g., tree heights, distances between trees and homes, sizes and/or masses of tree limbs, presence of ice dams, etc.) may be analyzed to determine a risk of damage to a structure or object located on the property. Based on the risk of damage, a risk output that includes (i) an indication of whether action should be taken to mitigate the risk of damage, (ii) an indication of whether insurance coverage associated with the structure or object should be offered, and/or (iii) a measure of the risk of damage, may be generated.

Abstract: Embodiments of a system and method for secure processing of image data are described. Image data included in image files may be encrypted and compressed in a single step using dynamically identified compression/encryption information such as code word tables.

Abstract: A computer-implemented method for targeting marketing content to an online user, includes the steps of collecting data describing the user or the user's ownership item in a user profile, comparing the user profile with a target profile and presenting the user with content based on the comparison. The user profile includes a photographic likeness of the user or the user's ownership item, as well as information derived from the photographic likeness. Certain content may be dynamically created for the user based on the user profile and that content may include the photographic likeness, which may be altered to incorporate a portion of the content into the photographic likeness. In this manner, personalized advertisements for cosmetics, jewelry, clothing, cosmetic procedures, home improvement products, and other products or services may be created. The user profile also includes data describing the user's browsing habits and responses to context-specific survey questions.

Abstract: An image processing apparatus divides a base image corresponding to a plurality of first images captured from different viewpoints, and a plurality of second images each corresponding to one or more of the plurality of first images, into a plurality of sub-bands including a sub-band that includes high-frequency components and a sub-band that does not include high-frequency components. As for the plurality of second images, the image processing apparatus then encodes the sub-band that includes high-frequency components. As for the base image, the image processing apparatus encodes the sub-band that does not include high-frequency components.

Abstract: A processing apparatus that enables margin correction of X-ray intensity data includes a calculation unit configured to calculate an expansion rate of X-ray intensity data detected in a unit region on a marginal side on a basis of an intensity distribution resulted from detection of uniform X-ray and a first function generation unit configured to generate an allocation function that allocates the X-ray intensity data detected in the unit region on the marginal side to an outside unit region, on the basis of the expansion rate. Enabling margin correction on the basis of the intensity distribution resulting from detection of uniform X-ray makes it possible to correct the distortion of intensity distribution on the marginal side. As a result, it is possible to expand the detection region of X-ray intensity data.

Abstract: A crash beam receiver for coupling to a crash beam for reducing the effects of a vehicle collision. The crash beam receiver includes a plurality of surfaces that are coupled together to form a hollow opening for receiving the crash beam. The crash beam receiver also includes a gate structure for providing a stopping point for the crash beam when the crash beam is inserted into the crash beam receiver. The gate structure includes a left gate, a right gate, and a protrusion structure that help to linearize the self-collapsing process of the crash beam in the longitudinal direction. The protrusion structure has a certain thickness such that the crash beam is caused to collapse inward toward its center in the event of a vehicle collision.

Abstract: Glioblastoma multiforme (GBM) is the most aggressive type of brain cancer. It is critical to determine the most effective patient-specific treatment quickly. Exemplary embodiments use a data-driven approach to extracting brain tumor information from data obtain from Whole Slide Image that is uploaded through an interface. Exemplary embodiments generate the following information about a glioblastoma tumor from a brain biopsy slide using neural networks: annotated areas of relevant tissues, molecular subtype, and expression status of an important gene and include three steps: the segmentation of tumor features; prediction of molecular subtype; and prediction of gene methylation status from a WSI.

Abstract: A watermarked content generating device including processing circuitry which performs active element conversion that converts a specified active modification target element from among elements included in an original content, based on a modification control code that is generated based on embedded information, when generating a structured content including at least a structured document, and performs passive element conversion that converts a specified passive modification target element from among elements included in the original content such that a change in a perceptual aspect of the original content corresponding to an active modification result element which results from the active element conversion is returned to having a same or equivalent perceptual aspect as that of the original content.

Abstract: In one or more embodiments, one or more methods, procedures, and/or systems may determine multiple most utilized colors of an image; may determine multiple lightness values respectively associated with the multiple most utilized colors of the image; may determine a lightness value of the multiple lightness values that is within a range of lightness values and a saturation value respectively associated with the lightness value that is above a threshold value; may determine a color of the multiple most utilized colors that is associated with the lightness value and with the saturation value as a dominant color; may repeatedly darken the dominant color by an amount until a ratio between a relative luminance value of a foreground color and a relative luminance value of the dominant color is within a range of a target ratio; and may utilize the dominant color as a background color.

Abstract: Systems and methods for providing data privacy in a private distributed ledger are disclosed. According to another embodiment a distributed ledger network may include a first node comprising a first node computer processor and hosting a central ledger comprising a plurality of entries for public transactions and private transactions, wherein the entries for public transactions comprise transaction payloads for the respective public transaction, and the entries for private transactions comprise a cryptographic hash digest of a transaction payload for the respective private transaction; and a plurality of second nodes each comprising a second node computer processor and hosting a public database comprising the public transactions, and a private database comprising transaction payloads for the private transactions to which the node is a party.

Abstract: A system for counting objects in a queue waiting for service at one of a plurality of electronic points of service “EPOSs” includes an object tracking system and a processor system for processing data. The processor system may have memory means storing data indicative of the positions of each of said EPOSs within a service point area, and be adapted to receive data from each EPOS indicating its operational status and to receive from said tracking system data relating to positions of objects within the service point area. The method of counting objects in a queue comprises firstly determining an initial seed location as a point at which a single queue of objects forms. Then, determining a new seed location as a point at which the single queue forms Once the new seed location is determined, tracked objects may be assigned to the single queue and then counted.

Abstract: Reusable network of spatially-multiplexed microfliuidic channels each including an inlet, an outlet, and a cuvette in-between. Individual channels may operationally share a main or common output channel defining the network output and optionally leading to a disposable storage volume. Alternatively, multiple channels are structured to individually lead to the storage volume. An individual cuvette is dimensioned to substantially prevent the formation of air-bubbles during the fluid sample flow through the cuvette and, therefore, to be fully filled and fully emptied. The overall channel network is configured to spatially lock the fluidic sample by pressing such sample with a second fluid against a closed to substantially immobilize it to prevent drifting due to the change in ambient conditions during the measurement. Thereafter, the fluidic sample is flushed through the now-opened valve with continually-applied pressure of the second fluid.

Abstract: Systems and methods for predicting vehicle behavior includes capturing images of a vehicle in traffic using an imaging device. Future behavior of the vehicle is stochastically modeled using a processing device including an energy-based model stored in a memory of the processing device. The energy-based model includes generating a distribution of possible future trajectories of the vehicle using a generator, sampling the distribution of possible future trajectories according to an energy value of each trajectory in the distribution of possible future trajectories an energy model to determine probable future trajectories, and optimizing parameters of each of the generator and the energy model using an optimizer. A user is audibly alerted with a speaker upon an alert system recognizing hazardous trajectories of the probable future trajectories.

Abstract: Systems, apparatuses and methods to provide image data, augmented with related data, to be displayed on a mobile computing device are disclosed. An example mobile device includes a camera to provide images of a scene from different angles to a server, at least one sensor to sense a position and an orientation of the camera, and a screen to present augmented reality data over the scene based on the position and the orientation of the camera and on a three-dimensional representation of the scene based on the images.

Abstract: Described in detail herein are methods and systems for performing a physical object assessment. The system entails an automated system to receive, unpack and assess a physical object based on automated determined attributes. The system can determine an element based on the assessment and generate a webpage including the attributes of the physical object and the determined element.

Abstract: Methods, apparatus, and articles of manufacture are disclosed to generate a synthetic point cloud of a spacecraft. An example apparatus includes a point cloud generator to generate a first synthetic point cloud of a first simulated space vehicle based on a simulated illumination source and a simulated image sensor, where the simulated illumination source and the simulated image sensor is operatively coupled to a second simulated space vehicle at a first position, where the simulated image sensor measures a parameter of the first simulated space vehicle, where the simulated illumination source uses a first configuration.

Abstract: Described is an, such as an unmanned aerial vehicle (“UAV”), that includes stereo pairs of imaging element, each imaging element including a region of interest controller. The region of interest controller for an imaging element of the stereo pair receives movement information affecting the imaging element and selects a portion of pixels of a digital image formed by the imaging element. The portion of pixels are provided to an image processor that utilizes the portion of pixels to determine depth information for objects represented by the pixels.

Abstract: A Key Tracking System and Method of Operation allows a user to store a digital image of a key in a memory storage on a mobile device. The digital image of the key is stored with a plurality of keys for a user, creating a database of key images that are viewable for comparison, cutting references, key location identity, and other common key organization and tracking features. The stored digital image of the key is used to comparatively identify the key amongst digital images of a plurality of keys. The stored digital image of key can be compared with an image of a corresponding recall key in order to cut a key bitting for the key that substantially matches the key bitting of the recall key. The stored digital image of the key is associated with a location of a corresponding keyway structure that graphically displays on a digital map.

Abstract: Systems for dynamically detecting duplicate items are provided. In some examples, a system, may receive a document, document image, document details or the like. The system may identify at least one other document for comparison to the received document. The system may perform optical character recognition on the document or document image to identify data within one or more standard fields of the document. The detected data may be compared to the at least one other document and, if no matches exist, the document may be identified as not a duplicate. If one or more matches exist, the system may identify x and y coordinates of a location of data within one or more standard fields on the received document. A corresponding location on the at least one other document may be identified based on the x and y coordinates.

Abstract: Image acquisition for meat grading. In one embodiment, a method includes aligning a suspended slaughtered animal half according to a predetermined orientation and retaining this orientation by a positioning device designed as part of a processing line until image acquisition of a cut surface; executing a 3D measurement and setting a trigger point in a control device, wherein the position of the cut surface is determined by an image processing device from a three-dimensional image recorded by a first image acquisition device; aligning the optical components of a second image acquisition device in accordance with the position in space determined for the cut surface; and trigger-initiated execution of at least one two-dimensional image acquisition of the cut surface by the second image acquisition device, the optical components of which are aligned beforehand, when a trigger condition coupled to the trigger point is met, as identified by the above control device.

Abstract: The disclosure provides a method of making watermarking data embedded in an Elementary Stream ES accessible to a receiver. The ES comprises video and/or audio data and is to be transmitted as a Packetized Elementary Stream PES in a Transport Stream TS such that the receiver can use the watermarking data to watermark the video and/or audio data in the ES without reconstructing the ES from the TS. The watermarking data identifies one or more watermarking locations in the video and/or audio data in the ES to be modified by the receiver in accordance with the watermarking data. The method comprises determining respective one or more watermarking TS locations in the TS corresponding to the one or more watermarking locations. The method further comprises embedding, in the ES, the watermarking data including watermarking location information identifying the one or more watermarking TS locations, or modifying watermarking data embedded in the ES by adding the watermarking location information.

Abstract: Current technologies detect the alertness of a user using frontal face. However, there are no techniques to track the position of face from a side angle. A method and system for face position tracking of a user and alerting the user is disclosed. The method includes capturing images using a monocular camera and identifying a user's face from the images and storing the user's face as reference face. The user's face is divided into two regions and one or more corner points are identified. A centroid is created in the two regions and the corner points are joined using virtual lines to create a dual flexible spider model. One or more new corner points are created on fulfillment of pre-defined conditions. Angle and magnitude of flow vectors is determined using one or more new corner points and alerts are given to the user based on the angle and magnitude.

Abstract: In some embodiments, an exemplary inventive computer-implemented method may include steps, performed by a processor, of: obtaining training real representations of a real subject; obtaining a training synthetic representation having a visual effect applied to a synthetic subject; training a first neural network and a second neural network by: presenting the first neural network with training real representation and candidate meta-parameters of latent variables for the visual effect to generate a training photorealistic-imitating synthetic representation of the real subject with the visual effect; presenting the second neural network with the training photorealistic-imitating synthetic representation and the training synthetic representation to determine actual meta-parameters of the latent variables of the visual effect, where the actual meta-parameters are meta-parameters at which the second neural network has identified that the training photorealistic-imitating synthetic representation is realistic, and pr

Abstract: An apparatus includes a plurality of optical filters, a stereo camera and a circuit. The optical filters may be configured to (i) generate polarized light by polarizing light received from a reflective surface and (ii) rotate in synchronization to each other. The stereo camera may be disposed adjacent the optical filters and may be configured to generate two sequences of images in electrical form from the polarized light. The circuit may be configured to (i) generate a sequence of disparity images from the two sequences of images and (ii) adjust the rotation of the optical filters in response to the sequence of disparity images.

Abstract: A method of providing Internet commercialization services to a content owner may include obtaining rights from an owner of Web page contents on a Web page to commercialize the Web page contents, obtaining an ad from an advertiser for placement on the Web page, the advertiser agreeing to pay for placement of the ad, paying for traffic to the Web page based on a bid to a traffic partner and sharing profits with the owner of the Web page contents, if any, from payments by the advertiser above payment to the traffic partner for the traffic. The content owner may be a publisher of content on the Internet. The traffic partner may be a search engine. The method may include automatically providing Internet commercialization services to an Internet publisher. Various aspects of the method may be modified in a closed loop fashion to optimize profits.

Abstract: A method for manufacturing a security document including a security pattern generation method includes the steps of selecting a bounding region, an image object, and a raster image or a function; cutting the image object into image object partitions; and repeating the image object as a pattern, preferably a lattice pattern, in the bounding region wherein each of the image object partitions of an image object is transformed depending on the position area of the image object in the lattice pattern and the rank of the image object partition.

Abstract: A system, method and computer-readable medium are provided to enable digital bank endorsement. A digital image of a back side of a check may be placed in a computer memory. Appropriate coordinates for a bank endorsement may be determined. A bank endorsement may be automatically generated. The digital image may then be electronically altered by overlaying, merging, or rendering text of the generated bank endorsement. A modified digital image may be combined with an image of the front side of the check and stored and/or exported to check clearing operations.

Abstract: An image processing method includes: a first obtaining step of obtaining data of an image; a second obtaining step of obtaining additional information; a third obtaining step of obtaining information related to a printable area of a print medium on which the image based on the data obtained in the first obtaining step is printed; determining step of determining a position where the obtained additional information is embedded in the image on the basis of the information obtained in the third obtaining step; and embedding step of embedding the additional information with reference to the determined position.

Abstract: An intelligent lighting system employs energy efficient outdoor lighting and intelligent sensor technology in cooperation with video analytics processing. The lighting system selectively illuminates outdoor spaces and identifies and evaluates events in a scene monitored by a video camera, thereby to facilitate proactive and appropriate security responses to those events. Selective use of advanced lighting fixtures may significantly reduce costs of lighting areas that are monitored by security systems such as streets, public parks, and parking lots, while simultaneously improving security, safety, and traffic control. Energy savings alone, for a properly designed system, are estimated at 50%-90% of current usage. When combined with remote monitoring, such systems may prevent accidents and criminal activity.

Abstract: Disclosed are an evaluation method and an evaluation device for a facial key point positioning result. In some embodiments, the evaluation method includes: acquiring a facial image and one or more positioning result coordinates of a key point of the facial image; performing a normalization process on the positioning result coordinate and an average facial model to obtain a normalized facial image; and extracting a facial feature value of the normalized facial image and calculating an evaluation result based on the facial feature value and a weight vector.

Abstract: Described herein is a method for implementing augmented reality associated to carrier means comprising a medium of a physical or digital type, wherein the method comprises in sequence the steps of incorporating an information content within an optical image of the medium, transferring the optical image that incorporates the information content onto the carrier means, decoding the information content, and reproducing the information content on an electronic device, wherein said step of reproducing enables the user to display the augmented reality.

Abstract: An online system customizes content items to be displayed on electronic displays. The electronic displays may be located at a venue with a large amount of foot traffic from users such as a park, airport, restaurant, shopping center, or airport. The system receives information from an electronic display describing users in vicinity of, or interacting with, the electronic display. For instance, the electronic display receives information from client devices of the users or by using sensors such as a camera that captures images of the users. Additionally, the system may receive interaction data generated by a touchscreen device of the electronic display. Based on the received information, the online system can select content items that are customized based on the interests of the users. The online system can also provide customized content items for display on a user's client device based on the user's interaction with an electronic display.

Abstract: Real time vehicle occupant emergency health data systems and methods that link a vehicle subsystem, a cloud subsystem, and a public health profile subsystem such that, in a triggering vehicle emergency event or the like, the health status of a previously-identified or unidentified vehicle occupant is captured in real time by an on-board camera and/or other sensor. This vehicle occupant identification and health status information are variously transmitted to the cloud, where the vehicle occupant identification is coupled with available public health profile information, in the case of a previously-identified vehicle occupant. Subsequently, the health status information and public health profile information are transmitted to an emergency responder or the like, thereby aiding in rendering medical assistance to an injured or sick vehicle occupant. The systems and methods of the present disclosure may also provide data analytics information that may be used later to improve vehicle safety.

Abstract: A vending machine recognition apparatus includes: a receiving unit configured to receive a captured image obtained by capturing an image of a vending machine; and a recognition unit configured to recognize merchandise from the captured image and recognize a price of the merchandise from a peripheral image region of an image region of the recognized merchandise.

Abstract: The present disclosure describes methods and systems for handling personal identifiable information. In some implementations, the information is collected during the authentication of identification (ID) documents. The system includes a one-way hashing function that converts unique personal identifiable information into a unique digest. The system can store the digest without having to store the personal identifiable information. Because the hashing function generates the same digest when given the same input, the digests can be used as anonymized identifiers in place of the personal identifiable information.

Abstract: A structural analysis computing device for determining structural characteristics of an object pictured in a three-dimensional (3D) image may be provided. The structural analysis computing device may include a memory, a user interface, an object sensor configured to capture the 3D image of the object, and at least one processor in communication with the memory and the object sensor. The processor may be configured to access the 3D image including the object, automatically determine a first plurality of measurements of the object from the 3D image, and display the 3D image on the user interface. The processor may be further configured to generate a data file including the 3D image and the first plurality of measurements, and store the data file within the memory. The processor may also be configured to transmit the data file to an insurance server computing device for generation of an associated insurance claim form.

Abstract: Debugging information for content rendered for display on a client device can be included in one or more watermarks inserted into the content. Performance, state, configuration, debugging, and other information can be encoded into a watermark and placed in rendered content in such a way that the information is imperceptible to most users. When a user reports a potential issue or problem, the user (or another source) can provide a version of the image frame corresponding to the issue, which enables the watermark to be located and the data extracted. A checksum or other value can be used to ensure that a complete and valid watermark was obtained. The data extracted from the watermark can then be used to attempt to identify and/or diagnose the potential issue without having to have access to the client device or communication with the user.

Abstract: Methods and systems for comparing content are provided. A recognizable object which appears in frames of a reference video file is recognized. A meta-feature of the recognizable object is determined. A first vector indicative of the meta-feature of the identified recognizable object in the reference video file is created. The meta-feature of the identified recognizable object which appears in a suspect video file is determined. A second vector indicative of the meta-feature of the identified recognizable object in the suspect video file is created. A correlation between the first vector and the second vector is determined. The correlation between the first vector and the second vector is indicative of a degree of confidence that the suspect video file is a copy of the reference video file.

Abstract: Similarity acquisition means calculates a similarity in each combination of an examination cross-sectional image and a reference cross-sectional image between examination volume data and reference volume data. Adjustment value acquisition means acquires an adjustment value of the similarity based on a relationship between the cross-sectional positions of examination cross-sectional images in two combinations and a relationship between the cross-sectional positions of reference cross-sectional images in the two combinations. Association means associates the examination cross-sectional image and the reference cross-sectional image with each other based on a sum of all the similarities and all the adjustment values.

Abstract: A hidden information detector for image files extracts N least significant bits from each of a first set of pixels of an image file, wherein N is an integer greater than or equal to 1. The detector then applies a mask to each of the extracted N least significant bits to form a second set of pixel values and determines a first probability as to whether the second set of pixels encodes a hidden image. Responsive to the first probability exceeding a first threshold, the detector determines a second probability as to whether the second set of pixels matches an image encoded in the first set of pixels. Responsive to a determination that the second probability is less than a second threshold, the detector performs a non-image classifier on the second set of pixels.