Abstract: An embodiment of the present disclosure takes the form of a method carried out by a perception-network device. The perception-network device provides a first perspective view of a scene to a first branch of a neural network, and generates a feature map via the first branch based on the first perspective view. The perception-network device augments the generated feature map with features of a complementary feature map generated by a second branch of the neural network provided with a second perspective view of the scene. The perception-network device generates a perception inference via the neural network based on the augmented feature map.

Abstract: A method of acquiring detection zone in image, comprises: sequentially acquiring a plurality of images associated with an image acquiring scene by a camera, computing a plurality of moving traces of a plurality of objects in the plurality of images by a computing device, performing a clustering procedure to the moving traces by the computing device to obtain a detection zone, and displaying the detection zone and another image by a display device, with said another image different from the plurality of images and associated with the image acquiring scene.

Abstract: A game-agnostic event detector can be used to automatically identify game events. Game-specific configuration data can be used to specify types of pre-processing to be performed on media for a game session, as well as types of detectors to be used to detect events for the game. Event data for detected events can be written to an event log in a form that is both human- and process-readable. The event data can be used for various purposes, such as to generate highlight videos or provide player performance feedback. The event data may be determined based upon output from detectors such as optical character recognition (OCR) engines, and the regions may be upscaled and binarized before OCR processing.

Abstract: An operation device of a convolutional neural network includes a first operation unit performing a convolution operation on an input feature map based on a predetermined mosaic type to output at least one pixel information of an output feature map, a prediction unit predicting the other pixel information of the output feature map as one of a skip pixel and an operation pixel based on the at least one pixel information, and a second operation unit skipping the convolution operation with respect to the sip pixel to process the skip pixel as an output of zero.

Abstract: Disclosed herein are systems, methods, and devices for detecting traffic lane violations. In one embodiment, a method for detecting a potential traffic violation is disclosed comprising bounding a vehicle detected from one or more video frames of a video in a vehicle bounding box. The vehicle can be detected and bounded using a first convolutional neural network. The method can also comprise bounding, using the one or more processors of the edge device, a plurality of lanes of a roadway detected from the one or more video frames in a plurality of polygons. The plurality of lanes can be detected and bounded using multiple heads of a multi-headed second convolutional neural network. The method can further comprise detecting a potential traffic violation based in part on an overlap of at least part of the vehicle bounding box and at least part of one of the polygons.

Abstract: An apparatus comprising a memory to store a first image captured by a first camera and a second image captured by a second camera; and a processor comprising circuitry, the processor to identify viewpoint information defining a view for a stitched image, the stitched image to be generated from a combination of a plurality of images comprising the first image and the second image; and based on the viewpoint information and a projection type, determine transformation parameters for at least one reference region associated with overlapping regions of the first image and the second image in order to: localize overlapping regions of the first image and the second image to determine stitching parameters; combine the first image and the second image using the stitching parameters; and assign data of the combined first image and the second image to the stitched image.

Abstract: Embodiments of this application disclose a position and attitude determining method. The method includes acquiring, by tracking a feature point of a first marked image, position and attitude parameters of an image captured by a camera; using a previous image of a first image as a second marked image in response to the previous image of the first image meeting a feature point tracking condition and the first image failing to meet the feature point tracking condition; acquiring, position and attitude parameters of the image captured by the camera relative to the second marked image; acquiring position and attitude parameters according to the position and attitude parameters of the image relative to the second marked image and position and attitude parameters of each marked image relative to a previous marked image; and determining a position and an attitude of the camera according to the position and attitude parameters.

Abstract: This disclosure relates to the use of “pseudo-images” to perform image recognition, e.g., to perform facial image recognition. In an embodiment, the pseudo-image is obtained by starting with a real world image and, after optional preprocessing, subjecting the image to a non-linear transformation that converts the image into a pseudo-image. While real world objects (or, more generally, real world patterns) may be perceivable in the starting image, they cannot be perceived in the pseudo-image. Image recognition takes place by comparing the pseudo-image with a library of known pseudo-images, i.e., image recognition takes place in pseudo-image space without a return to real world space. In this way, robust image recognition is achieved even for imperfect real world images, such as, real world images that have been degraded by noise, poor illumination, uneven lighting, and/or occlusion, e.g., the presence of glasses, scarves, or the like in the case of facial images.

Abstract: Disclosed is a device comprising: a light guide configured to receive incident light from one or more light sources; a cover layer with a sensing surface configured to receive an input object to be sensed; and a set of photodetectors. The light guide includes diffraction gratings configured to diffract the incident light as diffracted light, wherein the diffracted light exits the light guide and travels towards the sensing surface. The diffracted light is reflected from the sensing surface as reflected light. The reflected light is sensed by the set of photodetectors to form an image of the input object.

Abstract: A method of reducing false positives and identifying relevant true alerts in a video management system includes analyzing images to look for patterns indicating changes between subsequent images. When a pattern indicating changes between subsequent images is found, the video management system solicits from a user an indication of whether the pattern belongs to one of two or more predefined categories. The patterns indicating changes between subsequent images are saved for subsequent use. Subsequent images received from the video camera are analyzed to look for patterns indicating changes between subsequent images. When a pattern indicating changes between subsequent images is detected by the video management system, the video management system compares the pattern indicating changes between subsequent images to those previously categorized into one of the two or more predefined categories. Based on the comparison, the video management system may provide an alert to the user.

Abstract: An information processing apparatus including a setting processor that sets metadata for objects included in captured images captured in a time series by an imaging device based on the captured image, in which the setting processor processes the captured images in a reverse time series reverse to the time series from a predetermined event detected from the captured images and sets the metadata for objects involved in the predetermined event among the objects included in the captured images.

Abstract: When there is an image component common to multiple registered images, the accuracy of individual identification lowers. A decision unit decides, for each partial area, the degree of effectiveness relating to calculation of a score representing a similarity between a registered image obtained by capturing an object to be registered and a collated image obtained by capturing an object to be collated, based on whether or not the partial area contains an image component common to multiple registered images obtained by capturing multiple objects to be registered. A calculation unit calculates the score based on image components contained in the registered image and the collated image and on the degrees of effectiveness.

Abstract: The present embodiments relate to automatically estimating a three-dimensional pose of an object from an image captured using a camera with a structured light sensor. By way of introduction, the present embodiments described below include apparatuses and methods for training a system for and estimating a pose of an object from a test image. Training and test images are sampled to generate local image patches. Features are extracted from the local image patches to generate feature databased used to estimate nearest neighbor poses for each local image patch. The closest nearest neighbor pose to the test image is selected as the estimated three-dimensional pose.

Abstract: An electronic device and a method for controlling the electronic device, the method including receiving an input of a command to reproduce music contents, determining audio characteristics information on the music contents and situation information on an environment where the music contents are being reproduced; and displaying a visualization effect of visualizing the music contents using the audio characteristics information and the situation information, and reproducing the music contents.

Abstract: This disclosure relates to the use of “pseudo-images” to perform image recognition, e.g., to perform facial image recognition. In an embodiment, the pseudo-image is obtained by starting with a real world image and, after optional preprocessing, subjecting the image to a non-linear transformation that converts the image into a pseudo-image. While real world objects (or, more generally, real world patterns) may be perceivable in the starting image, they cannot be perceived in the pseudo-image. Image recognition takes place by comparing the pseudo-image with a library of known pseudo-images, i.e., image recognition takes place in pseudo-image space without a return to real world space. In this way, robust image recognition is achieved even for imperfect real world images, such as, real world images that have been degraded by noise, poor illumination, uneven lighting, and/or occlusion, e.g., the presence of glasses, scarves, or the like in the case of facial images.

Abstract: Methods and systems for finding and ranking video sequences that contain tactical sequences (e.g., an uninterrupted portion of a sporting event) that are similar to tactical event content in a target video sequence, such as those occurring in sports, include learning and labelling (e.g., through use of metadata and the like) tactically significant sequences in a plurality of video segments, including a plurality of such tactically significant sequences in a single video segment.

Abstract: A system for data-based pre-distortion for a nonlinear power amplifier includes a digital pre-distortion (DPD) component, including a DPD processor and a DPD calibration engine, where the DPD processor applies a set of DPD coefficients to a digital baseband data signal, to generate a pre-distorted digital baseband data signal for conversion to a radio frequency (RF) signal and amplification by a nonlinear power amplifier (PA) to generate an RF output signal, where the DPD calibration engine compares a digitized, down-converted version of the RF output signal with the digital baseband data signal, to determine distortion coefficients of the nonlinear PA, and to update the set of pre-distortion coefficients in the DPD processor to compensate for the distortion coefficients of the non-linear PA, where data transmission is uninterrupted by the transmission of non-data calibration signals.

Abstract: Abandoned articles left by a user departing from an autonomous vehicle are automatically detected by capturing image data including a plurality of diversely-illuminated images of a target area within a passenger cabin of the vehicle. A plurality of normal vectors are determined for respective pixels representing the target area in a normal extractor based on the images. A normal-driven map is stored in a first array in response to the plurality of normal vectors. A baseline map is stored in a second array compiled from baseline images of the target area in a nominal clean state. Differences between the normal-driven map and the baseline map indicative of an object not present in the clean state are detected in a comparator. Difficult to detect objects can be found using a single, fixed camera.

Abstract: Various embodiments of the present disclosure provide methods of making wellbore fluids with enhanced electrical conductivities. In some embodiments, such methods comprise: (1) pre-treating a carbon material with an acid; and (2) adding the carbon material to the wellbore fluid. Further embodiments of the present disclosure pertain to wellbore fluids formed by the methods of the present disclosure. Additional embodiments of the present disclosure pertain to methods for logging a subterranean well by utilizing the aforementioned wellbore fluids.

Abstract: In at least one embodiment, a reinforcement-learning-based searching approach is used to produce a training configuration for a machine-learning model. In at least one embodiment, 3D medical image segmentation is performed using learned image preprocessing parameters.