41results about How to "Robust identification" patented technology

In some examples, the technology identifies media and provides a user with supplemental information (e.g., supplemental media, a selectable link, etc.) based on the identity of the media. In other examples, the technology identifies media and provides a consumer with an option to click on a link associated with the media with a remote control to direct the video stream directly to a website sponsored by the commercial entity associated with the media. In other examples, the technology identifies media displayed on a subscriber's first computing device and displays the same media and / or a related media on the subscriber's second computing device.

The invention concerns an electronic device equipped with a video imaging process capability, which device includes a camera unit arranged to produce image frames from an imaging view which includes a region-of-interest ROI, an adjustable optics arranged in connection with the camera unit in order to focus the ROI on the camera unit, an identifier unit in order to identify a ROI from the image frame, a tracking unit in order to track the ROI from the image frames during the video imaging process and an auto-focus unit arranged to analyze the ROI on the basis of the tracking results provided by the tracking unit in order to adjust the optics. The device is arranged to determine the spatial position of the ROI in the produced image frame without any estimation measures.

A fundus camera using infrared light sources, which included an imaging optical path and an optical path for focusing and positioning and two optical paths share a common set of retina objective lens, and a computer-assisted method for retinal vessel segmentation in general and diabetic retinopathy (DR) image segmentation in particular. The method is primarily based on Multiscale Production of the Matched Filter (MPMF) scheme, which together the fundus camera, is useful for non-invasive diabetic retinopathy detection and monitoring.

The present invention relates to a method for recognizing a transverse parking space 32 and auxilarily parking in the transverse parking space 23. A parking assistance system provides a method of identifying a transverse parking space 32. The method also supports parking in the parking space 32. For this purpose, a travel motion of a motor vehicle 10 is evaluated by means of the following method steps: The travel motion 38 is evaluated for straight travel and in particular for portions 44 during which the yaw angle of the motor vehicle 10 remains constant. At least one edge 70, 72 or one corner E1, E2 of a first object 34 delimiting the transverse parking space 32 is detected, and an edge 70, 72 or a corner E1, E2 of a second object 36 delimiting the transverse parking space 32 is detected.

Signal egress from a shielding flaw in a cable telecommunication system is detected, even where signals carried by the cable telecommunication system are quadrature amplitude modulated signals that statistically resemble broadband noise by generating a marker signal comprising a double side band, suppressed carrier signal in the fringes of contiguous frequency bands and at a power level which cannot cause perceptible interference with signals in those contiguous frequency bands. The separation of the sidebands comprising the marker signal can unambiguously identify the marker signal and can distinguish between different cable telecommunication systems installed in the same geographic area. The marker signal can be additionally coded by varying the frequency and / or amplitude of the modulating signal used to create the marker signal.

Embodiments of the invention relate to a method for minimizing the influence of disturbing signals during calculation of shape elements from coordinate points. An aim of the embodiments of the invention is to exclude the coordinates which are not to be locally assigned to the desired shaped element from the calculation of the shaped element. Said aim is achieved by combining compensation methods for calculating the desired type of shaped element with recognition methods for the same type of shaped element and using the recognition methods for filtering the coordinate points that are relevant for calculating the shaped element out of all input coordinate points.

A device for organizing around the collapsible handle of wheeled luggage, baggage or containers the most personal, important, or frequently used items during travel is disclosed, which includes a bag that attaches to the collapsible handle of luggage and uses the handle as a skeleton frame to create a securing structure for those items without inhibiting the operation of the collapsible handle. The bag is fully accessible and convenient by its close proximity to the users hand when the handle is in its extended position. In either position, the device offers an increased opportunity for branding, personalization and identification for the end user. When the device is removed from the handle it can be stowed in a convenient configuration, such as a tri-fold, for storage or hung within convenient reach to the user to allow continued access to the items being stored.

According to embodiments of the present invention, machines, systems, computer-implemented methods and computer program products for mimicking a presence notification to optimize utilization of computing resources of a system comprising a plurality of virtual components and a virtual component manager are provided. Using the virtual component manager, one or more components that are each performing functions limited to providing a presence notification may be detected. The detected one or more components may be suspended, and the presence notifications for each suspended component may be generated and transmitted by the virtual component manager. Suspending a component that is solely providing a presence notification frees up resources that may be utilized for another component performing computational work.

The invention discloses a radar range profile statistics and recognition method based on a PPCA model in the strong noise background, which relates to the technical field of radar automatic target recognition and mainly solves the problem that the current statistics and recognition methods based on the PPCA model are not robust to noises. The training phase comprises the following steps: framing, translating, aligning and strength-normalizing radar HPPR continuously, learning the parameters of each azimuth frame of the PPCA model by adopting the processed HRRP and storing a template. The test phase comprises the following steps: first strength-normalizing, translating and aligning the samples to be tested and then estimating the range of the signal-to-noise ratios (SNR) of the samples; computing the distance value of each frame of each target and deciding the category attribute if the SNR is more than 30dB, and rewriting the distance value, solving the noise energy under SNR condition by minimizing the distance value, finally computing the distance value of each frame of each target and deciding the category attribute if the SNR is less than 30dB. The method has the advantages of robustness to noises and less computation and is applied to identifying radar targets.

The invention discloses a radar range profile statistics and recognition method based on a FA model in the strong noise background, which relates to the technical field of radar automatic target recognition and mainly solves the problem that the current statistics and recognition methods based on the FA model are not robust to noises. The training phase comprises the following steps: framing, translating, aligning and strength-normalizing radar HPPR continuously, learning the parameters of each azimuth frame of the FA model by adopting the processed HRRP and storing a template. The test phase comprises the following steps: first strength-normalizing, translating and aligning the samples to be tested and then estimating the range of the signal-to-noise ratios (SNR) of the samples; computing the distance value of each frame of each target and deciding the category attribute if the SNR is more than 30dB, and rewriting the distance value, solving the noise energy under SNR condition by minimizing the distance value, finally computing the distance value of each frame of each target and deciding the category attribute if the SNR is less than 30dB. The method has the advantages of robustness to noises and less computation and is applied to identifying radar targets.

The invention provides for a computer-implemented method for detecting one or more archive images matching a search image, each matching archive image being a derivative of the search image or being an original image the search image was derived from accessing a plurality of the archive images. For each of said archive images, a respective archive image histogram may be calculated, wherein each archive image histogram includes a plurality of combination micro-feature values. The archive image histogram may be stored to a database.

A method for dating photographic prints comprising scanning a front side of a photographic print forming an image scan; scanning a back side of the photographic print forming a back scan; defining a database of photographic print watermarks, each photographic print watermark comprising a representation of a watermark pattern and an associated date range; analyzing the back scan using a cross-correlation process between the back scan and the watermark patterns in the database of photographic print watermarks to identify a matching photographic print watermark; determining a date range for the image scan responsive to the identified matching photographic print watermark; and associating the determined date range with the image scan.

The invention provides a gain control method and a gain control device for audio data. The method in the embodiment of the invention comprises the following steps: acquiring Nth frame audio data and VAD information of the Nth frame audio data, and acquiring the expected gain of the Nth frame audio data according to an expected amplitude value and the Nth frame audio data; further acquiring the control gain of the Nth frame audio data according to the VAD information of the Nth frame audio data, VAD information of each frame of audio data in adjacent M frame audio data in front of the Nth frame audio data, the expected gain of the Nth frame audio data, and the expected gain of each frame of audio data in adjacent M frame audio data in front of the Nth frame audio data, so that the Nth frame audio data can be subjected to gain control processing by utilizing the control gain. Therefore, the amplitude value of the to-be-identified audio data is controlled to be above an identification threshold, and the voice recognition performance can be effectively prevented from being reduced.

The invention discloses a radar range profile statistics and recognition method based on a FA model in the strong noise background, which relates to the technical field of radar automatic target recognition and mainly solves the problem that the current statistics and recognition methods based on the FA model are not robust to noises. The training phase comprises the following steps: framing, translating, aligning and strength-normalizing radar HPPR continuously, learning the parameters of each azimuth frame of the FA model by adopting the processed HRRP and storing a template. The test phasecomprises the following steps: first strength-normalizing, translating and aligning the samples to be tested and then estimating the range of the signal-to-noise ratios (SNR) of the samples; computing the distance value of each frame of each target and deciding the category attribute if the SNR is more than 30dB, and rewriting the distance value, solving the noise energy under SNR condition by minimizing the distance value, finally computing the distance value of each frame of each target and deciding the category attribute if the SNR is less than 30dB. The method has the advantages of robustness to noises and less computation and is applied to identifying radar targets.

The invention provides a multi-attention feature fusion speaker recognition method, which comprises the following steps of: constructing a deep speaker characterization model which comprises a feature extractor and a speaker classifier, taking Fbank features as input of the deep speaker characterization model, extracting the Fbank features as speaker characterization through the feature extractor, in a training stage, using a speaker classifier to map the speaker representation to a speaker tag, and constructing a loss function to optimize a deep speaker representation model; and in a test stage, comparing similarity between speaker representations by adopting a cosine distance, and judging whether the speakers are the same speakers or not according to a threshold value. According to the method, weighted fusion is performed on the features of different branches through multi-attention feature fusion, and the multi-attention feature fusion comprises a space attention mechanism and a channel attention mechanism, so that effective information in each branch is enhanced, and speaker recognition performance with higher robustness is obtained.

Described herein are example methods and systems for neuro-behavioral relationships in dimensional geometric bedding (N-BRIDGE), which includes a comprehensive, data-driven analytic framework for mapping the multi-dimensional relationships between neural and behavioral features in humans N-BRIDGE allows mapping of variations along newly-defined data-driven behavioral dimensions that capture the geometry of behavioral / symptom variation to variation in specific neural features. A method for treating a patient based on neuro-behavioral mapping includes receiving, from a user interface of a computing device, behavioral data of a patient corresponding to mental health or cognitive status of the patient, predicting, by at least one processor of the computing device, a neural feature map for the patient representative of neural data based on the behavioral data, determining, by the at least one processor, a therapeutic associated with the neural feature map, and treating the patient with the therapeutic associated with the neural feature map.

The invention discloses an order address identification method, a system, a device, and a storage medium, and the method comprises the steps: segmenting a to-be-identified order address into a plurality of address lexical items according to an address hierarchy; respectively counting the number of target orders corresponding to each address lexical item; judging whether the number of the target orders corresponding to the address lexical items with the lower address hierarchy is greater than the number of the target orders corresponding to the address lexical items with the higher address hierarchy or not; and if so, determining that the order address is a malicious address. According to the malicious address recognition mode, the malicious address can be recognized more stably and comprehensively, a malicious address library does not need to be constructed, and the malicious address recognition mode can be suitable for any scene where the malicious address needs to be recognized.

Methods and system for identifying and evaluating on-street parking spots of an area are disclosed. The method comprises using at least one of satellite and aerial images and map data to identify vehicles, match them to street sections, compute on-street vehicle lanes, consolidate data from a plurality of images and identify parking lanes, as well as parking spots. The described system comprises a memory component, a processing component and an output component. Furthermore, a computer program executable by a computer and a non-transient computer-readable medium for identifying and evaluating on-street parking spots are described.

A method for identifying color in a target creates a ratio color space by determining the largest color component value of each pixel in an image and creating a ratio of all of the color component value with the largest component value for each pixel. The ratio for the color component of each pixel undergoes a threshold test to identify each color component as a rich shade or a fade shade. The ratio space color components are converted to a black and white image. Color information of adjacent pixels are clumped together to form blobs of the same color. The blobs are filtered by shape, color, location, or orientation and sorted to find targets that consist of a predefined pattern with the desired characteristics.

The invention provides a high myopia detection method based on an eye fundus image. The method comprises the following steps: acquiring a to-be-detected eye fundus image; inputting the fundus image to be detected into a high myopia detection model to obtain a high myopia detection result of the fundus image to be detected; wherein the high myopia detection model is obtained by training by using an adaptive course learning module and a semi-supervised transfer learning framework based on a general data set, a labeled fundus image sample data set and a label-free fundus image sample data set; wherein the adaptive course learning module is used for screening eye fundus image samples with labels and sequentially transmitting the screened eye fundus image samples with the labels to the student classification network model according to a determined sequence; the semi-supervised transfer learning framework comprises a pre-training source model and a semi-supervised teacher network model.

The invention provides a high-resolution range profile target robust recognition method for radar carrier frequency conversion, and mainly solves the problem that the radar target recognition rate is reduced due to mismatching of a test sample and a training sample under the carrier frequency conversion in the prior art. According to the implementation scheme, the method comprises the following steps: preprocessing radar high-resolution range profile data to obtain high-resolution range profile time-frequency domain feature data; establishing a time-frequency domain characteristic radar target database of the high-resolution range profile and setting a label value; selecting a training sample set and a test sample set before and after carrier frequency conversion from the radar target database; constructing a residual network; training the residual network under the original carrier frequency, and obtaining the residual network under the new carrier frequency through fine tuning; and inputting the test sample set under the new carrier frequency into the fine-tuned residual network under the new carrier frequency to obtain a target identification result. According to the method, the target identification performance under the radar carrier frequency conversion condition is improved, and the method can be used for robust identification of high-resolution range profile data of radar carrier frequency conversion.

The invention provides a face living body detection network training method and device, and the method comprises the steps: obtaining face images of a plurality of domains, each domain comprises a plurality of face images, and the plurality of face images in each domain are collected in a same environment scene and / or employing a same data collection device; selecting at least one domain from the plurality of domains as a source domain; inputting the face image in the source domain into an adversarial sample generation network to generate an adversarial sample image; and taking the face image in the source domain as training data, taking the adversarial sample image as test data, and training a to-be-trained face in-vivo detection network to obtain a trained face in-vivo detection network. According to the method and the device, a large number of adversarial samples can be obtained through the face images collected by various environment scenes and / or various types of data collection equipment, the adversarial samples are used for training the face in-vivo detection network, and the robustness of the trained face in-vivo detection network model is improved.

The invention relates to a disaggregation apparatus (1) for identifying an electrical consumer in an electrical network (2). An electrical signature providing unit (7) provides electrical signatures of the electrical consumers (4, 5, 6), and an electrical parameter determining unit (8) determines an overall electrical parameter of the electrical network (2). An identification unit (9) identifies an electrical consumer depending on the determined overall electrical parameter and a correlation of the electrical signatures. Since the identification unit identifies an electrical consumer depending on the determined overall electrical parameter and a correlation of the electrical signatures, the identification of an electrical consumer does not depend on the detection of an event only. This makes the identification more robust, especially less prone to errors caused by missed events, thereby improving the reliability of identifying an electrical consumer in the electrical network.

The invention provides a visual semantics and position sensing method and system based on infrared thermal imaging. The method comprises the steps: acquiring an infrared thermal image through an infrared thermal imager; performing processing classification and instance segmentation on the image through a deep neural network; obtaining pixel coordinates of the ROI by using examples such as a minimum external matrix and an image moment; calibrating a matrix transformation relation from an image coordinate system to a camera coordinate system by using a visual calibration algorithm; performing normal distance measurement through distance measuring equipment; determining the category of the object in the image and the three-dimensional coordinates of the ROI; and transmitting data to an uppercomputer in a distributed node mode. According to the invention, the influence caused by environmental uncertainty is reduced, and stable and uniform light supplementing equipment does not need to beindependently installed. In addition, three-dimensional positioning is conducted on the object in the mode that the ultrasonic sensor and the depth camera are combined, and the positioning accuracy and stability are improved. And a reliable technical method is provided for rapid, accurate and robust visual identification and positioning.