Abstract: A semiconductor device for an apparatus having a movement drive unit and an image unit, includes an image detection unit, an image recognition unit and control unit. The image detection unit detects an object in a captured image and cuts out an image area including the object from the captured image as an object detection area image. The image recognition unit performs an image recognition processing for the object detection image area and output a recognition probability of the object. The control unit controls at least one of a moving speed of the movement drive unit and an imaging interval of the image unit based on the recognition probability.

Abstract: This facial authentication device is provided with: a detecting means for detecting a plurality of facial feature point candidates, using a plurality of different techniques, for at least one facial feature point of a target face, from a plurality of facial images containing the target face; a reliability calculating means for calculating a reliability of each facial image, from statistical information obtained on the basis of the plurality of detected facial feature point candidates; and a selecting means for selecting a facial image to be used for authentication of the target face, from among the plurality of facial images, on the basis of the calculated reliabilities.

Abstract: In various examples, detected object data representative of locations of detected objects in a field of view may be determined. One or more clusters of the detected objects may be generated based at least in part on the locations and features of the cluster may be determined for use as inputs to a machine learning model(s). A confidence score, computed by the machine learning model(s) based at least in part on the inputs, may be received, where the confidence score may be representative of a probability that the cluster corresponds to an object depicted at least partially in the field of view. Further examples provide approaches for determining ground truth data for training object detectors, such as for determining coverage values for ground truth objects using associated shapes, and for determining soft coverage values for ground truth objects.

Abstract: A method for training an artificial neural network on an additional untrained segmentation task prevents the loss of previously acquired segmentation skills on originally trained segmentation tasks.

Abstract: The present invention discloses a system for selection of candidate video frames from traffic video comprising an imaging processor operatively connected to traffic video source. The imaging processor receives the traffic video from said source for selection of candidate video frames from the received traffic video for analysis of entire traffic video stream to extract vehicles details from therefrom including detection of situations of moving vehicles and non-moving vehicles. The traffic video is forwarded to the imaging processor from the source at an appropriate rate suitable for said candidate frames selection computation.

Abstract: A method for real-time detection and correction of shadowing in hyperspectral retinal images may include capturing receiving, using a processor, a hyperspectral image of a retina of a patient, detecting, by the processor, a shadow in the hyperspectral image, determining, by the processor that the shadow of the hyperspectral image exceeds a threshold, and in response to determining that the shadow of the hyperspectral image exceeds the threshold, initiating, using the processor, a capture of an additional hyperspectral image of the retina of the patient. Various other methods, systems, and computer-readable media are also disclosed.

Abstract: A set of gesture commands are learned. Each gesture in the set of gesture commands controls a function of one or more computing devices. A first gesture received from a user is identified. A determination is made whether at least one computing device associated with the identified gesture is active. In response to determining that at least one computing device associated with the identified gesture is active, a determination is made whether the user associated with the identified gesture is recognized. In response to determining that the user associated with the identified gesture is recognized, an action to perform on at least one computing device associated with the identified gesture is determined. The determined action is implemented.

Abstract: Disclosed herein are various methods of calibrating relative orientation between spatial orientation sensors, relative orientation between imaging sensor and spatial orientation sensor, and relative orientation between imaging sensors. The relative orientation of the spatial orientation sensors, imaging sensors is calculated with respect to a reference frame.

Abstract: Embodiments of the disclosure disclose a method for automatically annotating a target object in images. In one embodiment, the method comprises: obtaining an image training sample including a plurality of images, wherein each image of the plurality of images is obtained by photographing a same target object, and the adjacent images share one or more same environmental feature points; using one of the plurality of images as a reference image to determine a reference coordinate system, and create a three-dimensional space model based on the three-dimensional reference coordinate system; determining the position information of the target object in the three-dimensional reference coordinate system upon the three-dimensional space model being moved to the position of the target object in the reference image; and mapping the three-dimensional space model to image planes of each image, respectively, based on respective camera pose information determined based on environmental feature points in each image.

Abstract: Disclosed herein are system, method, and computer program product embodiments for analyzing contextual symbol information for document processing. In an embodiment, a language model system may generate a vector grid that incorporates contextual document information. The language model system may receive a document file and identify symbols of the document file to generate a symbol grid. The language model system may also identify position parameters corresponding to each of the symbols. The language model system may then analyze the symbols using an embedding function and neighboring symbols to determine contextual vector values corresponding to each of the symbols. The language model system may then generate a vector grid mapping the contextual vector values using the position parameters. The contextual information from the vector grid may provide increase document processing accuracy as well as faster processing convergence.

Abstract: The systems and methods disclosed herein are generally directed to the real-time symbol rate of modulation modification based on noise and interference detection on a received signal during data transmission across a network. In one embodiment, a receiving device can measure the power of affected carriers, subcarriers, and/or tones. In one embodiment, adaptive bit loading can be used such that a lower order modulation scheme can be used to overcome through the noise and increase the Signal-to-noise ratio (SNR) of the transmitted signal. In one embodiment, more throughput (for example, more bps/Hz) can be achieved in the network by using subcarriers that have a higher SNR. Further, the bit-loading can serve to maximize the data rate subject to power and bit-error ratio (BER) constraints of the network.

Abstract: A cable modem transceiver includes a processor configured to derive an instant of time for an upstream calibration signal on basis of upstream scheduling information. Further, the cable modem transceiver includes a transmitter configured to generate the upstream calibration signal at the derived instant of time. The cable modem transceiver additionally includes a detector configured to determine a property of the generated upstream calibration signal. The processor is further configured to derive at least one calibration parameter for the transmitter on basis of the detected property.

Abstract: Coverage enhancements and coverage mode switching related optimizations are discussed for user equipments (UEs) that may switch between various coverage extension (CE) and non-CE modes of operation. In such enhancements, paging uncertainty and delays may be reduced by sending pages either simultaneously or using historical information over multiple coverage modes available to UEs. Random access procedures may be improved by providing CE mode random access procedures that are available when normal mode random access attempts fail and before declaring radio link failure. Additional aspects include improvements for more advanced UEs to improve coverage within normal mode operations by leveraging techniques used for narrowband CE mode operations, including transmission repetition and gapless transmission scheduling over hopped narrowband frequencies.

Abstract: This application discloses a communication method, a communications apparatus, a network device, and a terminal. Quasi co-location QCL beam information is determined. The QCL beam information includes beam information that has a QCL relationship with a reference signal antenna port. The QCL beam information is sent. A terminal receives the QCL beam information. The QCL beam information includes the beam information that has the QCL relationship with the reference signal antenna port. The terminal determines, based on the beam information that is included in the QCL beam information and that has the QCL relationship with the reference signal antenna port, a beam that has the QCL relationship with the reference signal antenna port.

Abstract: A transmitter for providing channel equalization that includes a first driver and second driver having a high pass filter. The first driver generates a first output signal representing a digital input signal. The second driver generates a second output signal representing a high pass filtered version of the digital input signal. The first and second output signals are summed to provide a third output signal that is channel equalized for transmission over a channel.

Abstract: Techniques are disclosed relating to rendering graphics objects. In some embodiments, a graphics unit is configured to transform graphics objects from a virtual space into a second space according to different transformation parameters for different portions of the second space. This may result in sampling different portions of the virtual space at different sample rates, which may reduce the number of samples required in various stages of the rendering process. In the disclosed techniques, transformation may occur prior to rasterization and shading, which may further reduce computation and power consumption in a graphics unit, improve image quality as displayed to a user, and/or reduce bandwidth usage or latency of video content on a network. In some embodiments, a transformed image may be viewed through a distortion-compensating lens or resampled prior to display.

Abstract: A method and apparatus obtains a source image having a plurality of source color gamut pixels in a source color gamut. The method and apparatus converts the plurality of source color gamut pixels to a plurality of corresponding target color gamut pixels using non-linear interpolation of a plurality of output pixel values from a reduced 3-D look-up table (LUT) for a target color gamut. The method and apparatus provides, for display, the plurality of target color gamut pixels (e.g., one or more pixels) on a target color gamut display.

Abstract: A person trend recording device includes: an image acquisition unit configured to acquire moving image data obtained by capturing an image of a target area; a person detector configured to detect and track a person, who acts within the target area, by analyzing the moving image data; an event detector configured to detect an event that occurs with the person during a period of tracking the person, by analyzing the moving image data; a physio-psychological index detector configured to detect a temporal change in a physio-psychological index of the person during the tracking period, by analyzing the moving image data; and a trend record generator configured to generate, as trend record data on the person, data in which at least one or more events that occur with the person are temporally associated with the temporal change in the physio-psychological index of the person.

Abstract: A system for generating simulated body parts for images may include a body part recognition convolutional neural network (CNN) to recognize a body part in an input image. The body part recognition CNN may be trained using first training data including training images including body parts contained in the input image being identified. The system may also include a body part generative adversarial network (GAN) to complete an image of the body part in the input image based on a body part identification output by the body part recognition CNN. The body part GAN may be trained using second training data including at least partial training images.

Abstract: This application provides example beamforming (BF) training methods, receiving devices, and sending devices. One example method includes performing BF training on at least one channel with a first device based on BF training request information, where the BF training request information includes antenna configuration information of the BF training and channel configuration information of the at least one channel. First feedback information sent by the first device is received, where the first feedback information includes a measurement result of the BF training, information about an antenna corresponding to the measurement result, beam information of the antenna, and channel information corresponding to the antenna. At least one of an optimal antenna configuration or digital domain BF precoding information on the at least one channel is determined based on the first feedback information.