Abstract: A virtual reality (VR) sickness assessment method according to an embodiment includes receiving virtual reality content, and quantitatively evaluating virtual reality sickness for the received virtual reality content using a neural network based on a pre-trained neural mismatch model. The evaluating of the virtual reality sickness may include predicting an expected visual signal for an input visual signal of the virtual reality content based on the neural mismatch model, extracting a neural mismatch feature between the predicted expected visual signal based on the neural mismatch model and an input visual signal for a corresponding frame of the virtual reality content corresponding to the expected visual signal, and evaluating a level of the virtual reality sickness based on the neural mismatch model and the extracted neural mismatch feature.

Abstract: A Computer Vision (CV) model generated by a Machine Learning (ML) system may be retrained for more accurate computer image analysis in Robotic Process Automation (RPA). A designer application may receive a selection of a misidentified or non-identified graphical component in an image form a user, determine representative data of an area of the image that includes the selection, and transmit the representative data and the image to an image database. A reviewer may execute the CV model, or cause the CV model to be executed, to confirm that the error exists, and if so, send the image and a correct label to an ML system for retraining. While the CV model is being retrained, an alternative image recognition model may be used to identify the misidentified or non-identified graphical component.

Abstract: The subject technology receives a representation of a neural network (NN) model to be executed on an electronic device, the representation of the NN model including nodes corresponding to intermediate layers of the NN model. The subject technology determines, for the respective operation corresponding to each node in each respective intermediate layer of the NN model, a respective set of operations that are mathematically equivalent to the respective operation such that an aggregation of outputs of the respective set of operations is equivalent to an output of the respective operation. The subject technology generates a graph based on each respective set of operations, wherein the graph includes a set of branches, each branch includes a plurality of operations. The subject technology determines a respective order for executing each branch of the graph.

Abstract: Systems and methods for multi-directional imaging in coupling with an indentation device are provided. The systems and methods can enable in situ multi-directional and full-field interfacial force response on soft matter. With more than two two-dimensional (2D) plane videos and/or images, the in situ nano-, micro-, and meso-scale time-dependent damage responses at the three-dimensional (3D) level can be constructed. Based on the obtained 2D and 3D videos and/or images, the underlying dynamic force response and fatigue mechanisms can be observed using a digital image correlation technique.

Abstract: There is disclosed a computer-implemented method for lossy image or video compression, transmission and decoding, the method including the steps of: (i) receiving an input image at a first computer system; ({umlaut over (?)}) encoding the input image using a first trained neural network, using the first computer system, to produce a latent representation; (iii) quantizing the latent representation using the first computer system to produce a quantized latent; (iv) entropy encoding the quantized latent into a bitstream, using the first computer system; (v) transmitting the bitstream to a second computer system; (vi) the second computer system entropy decoding the bitstream to produce the quantized latent; (vii) the second computer system using a second trained neural network to produce an output image from the quantized latent, wherein the output image is an approximation of the input image.

Abstract: A method for video decoding includes decoding prediction information of a current block, the prediction information indicating a sub-block based inter prediction mode, the current block including a plurality of sub-blocks arranged into one or more areas. For each area, the method includes determining base motion vectors of a portion of the sub-blocks within the area, the portion including at least one non-corner sub-block of the current block, and converting the base motion vectors of the portion of the sub-blocks into respective target motion vectors.

Abstract: A system for providing golf information, the system which includes a green photographing unit for photographing an image of an area including a green on a golf course and communicatively connected to a communication network to transmit green image data to an outside, a golfer terminal that is carried by a golfer on the golf course and communicatively connected to a communication network to transmit golfer position data to the outside, and an information providing server that is communicatively connected to the communication network to generate golf information using the green image data and the golfer position data and transmit the generated golf information to the golfer terminal.

Abstract: A computer-aided method for training an artificial neuronal network for the simulation of an environment sensor of a driver assistance system includes the following work steps: reading in traffic scenarios of a test journey; deriving test journey data and the sensor data to be output by the environment sensor from the traffic scenarios; generating a data stream which depicts the traffic scenarios from the perspective of the environment sensor; outputting the data stream such that the environment sensor can generate sensor data on the basis of the data stream and can provide same to a data interface at which the test journey data and the sensor data to be output are also provided; and reading the provided data into the artificial neuronal network.

Abstract: A training device and a training method for a neural network model. The training method includes: obtaining a data set; completing, according to the data set, a plurality of artificial intelligence (AI) model trainings to generate a plurality of models corresponding to the plurality of AI model trainings respectively; selecting, according to a first constraint, a first model set from the plurality of models; and selecting, according to a second constraint, the neural network model from the first model set.

Abstract: A multi-camera positioning and dispatching system includes a plurality of cameras and processing device. The cameras are disturbed over an indoor space having a plurality of areas; the cameras are corresponding to the areas and capture the images of the areas respectively. The processing device converts the pixel coordinates of the image of the camera corresponding to each area into the camera coordinates of the area, and converts the cameras coordinates of the area into the world coordinates of the area so as to integrate the images with one another and obtain a panoramic map, defined by a world coordinate system, of the indoor space. The processing device projects the working unit in the image captured by any one of the cameras to the panoramic map. The system can achieve positioning function via the panoramic map so as to optimize indoor environment management and save manpower.

Abstract: An image processing apparatus comprises an image obtaining unit that obtains a plurality of images based on image capturing performed by a plurality of image capture apparatuses that perform image capturing of an imaging region from directions that are different to each other, and a panoramic image generating unit that generates a panoramic image with a specific position in the imaging region being a reference, based on a plurality of images obtained by the image obtaining unit.

Abstract: A method comprises obtaining input data comprising a hologram of a 3-dimensional (3D) particle field, a depth map of the 3D particle field, and a maximum phase projection of the 3D particle field. The method also comprises applying a U-net convolutional neural network (CNN) to the input data to generate output data. Encoder blocks have residual connections between a first layer and a second layer that skips over a convolution layer of the encoder block. Decoder blocks have residual connections between a first layer and a second layer that skips over a convolution layer of the decoder block. The output data includes a channel in which pixel intensity corresponds to relative depth of particles in the 3D particle field and an output image indicating locations of centroids of the particles in the 3D particle field.

Abstract: According to the present disclosure, an image decoding method performed by a decoding apparatus comprises the steps of: acquiring motion prediction information on a current block from a bitstream; generating an affine MVP candidate list including affine motion vector predictor candidates for the current block; deriving CPMVPs for CPs of the current block on the basis of one affine MVP candidate among the affine MVP candidates included in the affine MVP candidate list; deriving CPMVDs for the CPs of the current block on the basis of the motion prediction information; deriving CPMVs for the CPs of the current block on the basis of the CPMVPs and the CPMVDs; and deriving prediction samples for the current block on the basis of the CPMVs.

Abstract: A video decoding method includes determining whether an ultimate motion vector expression (UMVE) mode is allowed for an upper data unit including a current block, when the UMVE mode is allowed for the upper data unit, determining whether the UMVE mode is applied to the current block, when the UMVE mode is applied to the current block, determining a base motion vector of the current block, determining a correction distance and a correction direction for correction of the base motion vector, determining a motion vector of the current block by correcting the base motion vector according to the correction distance and the correction direction, and reconstructing the current block based on the motion vector of the current block.

Abstract: An artificial neural network-based method for detecting a surface type of an object includes: receiving a plurality of object images, wherein a plurality of spectra of the plurality of object images are different from one another and each of the object images has one of the spectra; transforming each object image into a matrix, wherein the matrix has a channel value that represents the spectrum of the corresponding object image; and executing a deep learning program by using the matrices to build a predictive model for identifying a target surface type of the object. Accordingly, the speed of identifying the target surface type of the object is increased, further improving the product yield of the object.

Abstract: A trackable camera beacon is provided that is mountable onto a camera, so that the camera can be more easily tracked and automatically controlled. The camera beacon obtains lens data from the lens of a camera, position data corresponding to the camera beacon, and outputs a unified and synchronized data packet. The unified and synchronized output includes position, orientation and lens data. This can be used to also control the camera, such as the focus, iris and zoom parameters of the lens.

Abstract: In a method for imaging a container holding a sample, the container is illuminated with a laser sheet that impinges upon the container in a first direction corresponding to a first axis. A plane of the laser sheet is defined by the first axis and a second axis orthogonal to the first axis. The method also includes capturing, by a camera having an imaging axis that is substantially orthogonal to at least the first axis, an image of the container. The method further includes analyzing, by one or more processors, the image of the container to detect particles within, and/or on an exterior surface of, the container.

Abstract: An arrangement for mounting an inside rearview mirror in a vehicle. The inside rearview mirror has a mirror head that is adjustably mounted, by the agency of a mirror base, on a base plate permanently attached to the vehicle. The mirror base is mounted on the base plate through a first support point so as to be pivotable about a vehicle transverse axis and also through a second, detachable rear support point. The first support point is formed by two bearing side pieces of the base plate. Each of the bearing side pieces has a bearing eye in which the mirror base is mounted by means of bearing axle(s) so as to be pivotable about a vehicle transverse axis. At least one bearing side piece is designed to be slotted toward and transversely to the central axis of the bearing eye for installation of the relevant bearing axle.

Abstract: Methods and systems involve obtaining information from one or more sources to determine a real-time context. A method includes determining a situational awareness score indicating a level of vigilance required based on the real-time context, obtaining images of eyes of a driver to detect a gaze pattern of the driver, determining a relevant attention score indicating a level of engagement of the driver in the real-time context based on a match between the gaze pattern of the driver and the real-time context, and obtaining images of the driver to detect behavior of the driver. A driver readiness score is determined and indicates a level of readiness of the driver to resume driving the vehicle based on the behavior of the driver. An off-road glance time is obtained based on using the situational awareness score, the relevant attention score, and the driver readiness score.

Abstract: An apparatus and a method use a convolutional neural network (CNN) including a plurality of convolution layers in the field of artificial intelligence (AI) systems and applications thereof. A computing apparatus using a CNN including a plurality of convolution layers includes a memory storing one or more instructions; and one or more processors configured to execute the one or more instructions stored in the memory to obtain input data; identify a filter for performing a convolution operation with respect to the input data, on one of the plurality of convolution layers; identify a plurality of sub-filters corresponding to different filtering regions within the filter; provide a plurality of feature maps based on the plurality of sub-filters; and obtain output data, based on the plurality of feature maps.