Abstract: The present disclosure provides a method and a device for training a neural network model for use in analyzing captured images, and an intelligent image capturing apparatus employing the same. The neural network model can be trained by performing the image reconstruction and the image classification using based on image data received from a plurality of image capturing devices installed in the monitoring area, calculating at least one loss function based on data processed by the neural network model or the neural network model training device, and determining parameters minimizing the loss function. In addition, the neural network model can be updated through the re-training taking into account the newly acquired image data. Accordingly, the image analysis neural network model can operate with high precision and accuracy.

Abstract: An object tracking system that includes a plurality of sensors and a tracking system. A first sensor from the plurality of sensors is configured to capture a first frame of a global plane for at least a portion of the space. The tracking system is configured to determine a pixel location in the first frame for an object located in the space, and to apply a homography to the pixel location to determine a coordinate in the global plane. The homography is configured to translate between pixel locations in the first frame and coordinates in the global plane.

Abstract: An information processing apparatus learns an attribute-positional relation in a form for learning to create an attribute-positional relation rule for a character string in a form, determines correspondence between a character string in a result of character recognition executed on an image of the form for learning and an attribute in form data for learning based on the form for learning to create a character string attribute model for acquiring a probability of an attribute of the character string in the form, applies the character string attribute model to a character string in a result of character recognition executed on an image of the form to acquire the probability of an attribute, and corrects the probability based on a position in the form of the character string in the result of the character recognition executed on the image of the form and the attribute-positional relation rule.

Abstract: An integrative computational soil mapping system and process that reduces the required number of soil property measurements without jeopardizing the statistical precision of the resulting digital soil maps. The integrative computational soil mapping system and process saves monetary resources and time by reducing the number of soil property measurements required to produce digital soil maps and by offering soil sample locations which capture the maximum amount of representativeness of the soil characteristics in a determined area. In addition, the inventive system and process are integrative computational soil mapping that utilize algorithms based on state-of-the-art computational statistics and machine learning methods for the production of digital soil property maps and also provides soil sampling locations to collect new soil property measurements.

Abstract: Method and system for extracting metadata from an observed scene comprising the use of high-speed image sensor array observing the scene through a field-of-view, filtering, matching of interesting information defined as activity/incident in the observed scene with a predefined-library, adapting the filtering to enhance frequencies of interest and suppress all others, and extracting metadata of interest.

Abstract: Provided is a neural network device including at least one processor configured to implement an arithmetic circuit configured to generate third data including a plurality of pixels based on a neural network configured to perform an arithmetic operation on first data and second data, and a compressor configured to generate compressed data by compressing the third data, wherein the compressor is further configured to generate, as the compressed data, bitmap data comprising location information about a non-zero pixel having a non-zero data value among the plurality of pixels based on a quad-tree structure.

Abstract: A system and method for error estimation in an eye gaze tracking system in a vehicle may include an operator monitoring system providing measured eye gaze information corresponding to an object outside the vehicle and an external object monitoring system providing theoretical eye gaze information and an error in the measured eye gaze information based upon the measured eye gaze information and the theoretical eye gaze information.

Abstract: A monitoring system for obtaining a video related to medication adherence of a user, includes: a wireless communication device comprising a motion sensor, an ambient light sensor, a first transceiver transmitting data to an external device, and a first controller configured to control the motion sensor, the ambient light sensor, and the first transceiver, the wireless communication device having an attaching portion for being attached to an object containing a medication; and a wearable device including a camera, a second transceiver receiving a signal from the first transceiver, and a second controller configured to obtain video data through the camera based on the signal received through the second transceiver.

Abstract: A method for random access for beam failure recovery. In the method, a random access configuration for the beam failure recovery is received. In the event of a beam failure, a random access procedure is performed according to the random access configuration.

Abstract: A vehicle parking assistance device includes an image sensing device, an artificial intelligence learning device, and a controller connected with the image sensing device and the artificial intelligence learning device. The controller is configured to obtain an image using the image sensing device, detect at least one parking line pair in the obtained image, detect a parking slot based on deep learning, detect a parking area based on the detected parking slot and the at least one detected parking pair, detect an entrance point for the parking area, and generate parking information based on the parking area and the entrance point.

Abstract: The present disclosure provides an integrated circuit chip device and a related product. The integrated circuit chip device includes: a primary processing circuit and a plurality of basic processing circuits. The primary processing circuit or at least one of the plurality of basic processing circuits includes the compression mapping circuits configured to perform compression on each data of a neural network operation. The technical solution provided by the present disclosure has the advantages of a small amount of computations and low power consumption.

Abstract: A system, method, and computer program product for implementing video action recognition is provided. The method includes receiving a video stream comprising user movement actions. Skeleton points associated with a video representation of a user executing the user movement actions are extracted and categorized with respect to multiple digital levels. Initial visual windows points are generated within video frames and an average movement distance for the group of skeleton points are determined with respect to the video frames. In response, sizes for the visual windows are adjusted and feature vectors are extracted from the group of skeleton points. Point coordinates of the skeleton points are extracted and linked with the feature vectors. A convolutional neural network associated with linking the feature vectors with the point coordinates is generated and the video stream is enabled with respect to video action recognition associated with accurate presentation of the video stream.

Abstract: A method for classifying objects which comprises a provision of measuring data from a sensor for a feature extraction unit as well as extraction of modality-independent features from the measuring data by means of the feature extraction unit, wherein the modality-independent features are independent of a sensor modality of the sensor, so that a conclusion to the sensor modality of the sensor is not possible from the modality-independent features.

Abstract: A method includes providing a computer system, the computer system including at least a processor and a memory, the memory including at least an operating system, executing a process in the memory, the process including providing a recurrent circuit model, and converting the recurrent circuit model into a recurrent neural network that can be fit with gradient descent.

Abstract: State of the art techniques in the domain of video analysis have limitations in terms of capability to capture spatio-temporal representation. This limitation in turn affects interpretation of video data. The disclosure herein generally relates to video analysis, and, more particularly, to a method and system for video analysis to capture spatio-temporal representation for video reconstruction and analysis. The method presents different architecture variations using three main deep network components: 2D convolution units, 3D convolution units and long short-term memory (LSTM) units for video reconstruction and analysis. These variations are trained for learning the spatio-temporal representation of the videos in order to generate a pre-trained video analysis module. By understanding the advantages and disadvantages of different architectural configurations, a novel architecture is designed for video reconstruction.

Abstract: A method for operating an artificial neural network (ANN) on a hardware platform. The ANN is designed to ascertain confidences with which input data are to be assigned to N discrete classes. The hardware platform includes a dedicated unit which forms from a list of M>N confidences expanded confidences by encoding into each confidence an identification number of its place in the list, and numerically sorts the expanded confidences. The unit is fed confidences 1, . . . , M?N, which have the minimal representable value, and confidences M?N+1, . . . , M, which correspond to the N discrete classes, and/or it is ensured that those confidences fed to the unit that correspond to one of the N discrete classes have a value higher than the minimal representable value. A ranking of the classes ordered according to confidences, to which the input data are to be assigned, is ascertained from the first N of the numerically sorted expanded confidences.

Abstract: A processing apparatus (10) including: an event extraction unit (11) that extracts, for each image of a time-series image group, an event in which a person and an object relate to each other; an event information generation unit (12) that generates event information indicating a time-series change of a first event defined by a first person, a first object, and a first relationship between a person and an object and also indicating a time-series change of a second event defined by a second person, a second object, and a second relationship between a person and an object; a causal relationship computation unit (13) that computes a score indicating a causal relationship between the first event and the second event; and an output unit (14) that outputs information indicating the first event and the second event for which the score satisfies a predetermined condition.

Abstract: An inpainting method includes obtaining image information at an electronic device, where the image information identifies an area corresponding to a removed object within an image. The method also includes reconstructing the area corresponding to the removed object by (i) applying a semantic mask and a surface normal map to identify and rank neighboring contexts of the area and (ii) sampling, using an attention mechanism, the ranked contexts to generate pixel information for the area. The method further includes rendering the image with the reconstructed area.

Abstract: An information processing apparatus includes one or more processors that function as the following units: a detection unit configured to detect a defect occurring in a structure by using local information about the structure based on a first image of the structure; a first determination unit configured to determine a global state of the structure based on a result of detection of the defect by the detection unit and global information indicating information about a wider area than an area of the structure indicated by the local information, the global information being based on a second image having a lower resolution than that of the first image; and a second determination unit configured to determine a degree of damage to a predetermined area of the structure based on a result of determination made by the first determination unit of the global state of the structure.

Abstract: An abnormality detection device is provided to detect an abnormal posture of a driver of a vehicle based on information of an image captured by an imaging unit for which an imaging region is defined to image a head of the driver. The abnormality detection device is configured to define at least one premonition region in a range in which the head is supposed to be positioned in process of transitioning from a normal posture to the abnormal posture in the imaging region. When the head stays in a premonition region, the abnormality detection device determines that the driver is in a state of the abnormal posture.