Abstract: Introduced here is a technique to create small compressed image files while preserving data quality upon decompression. Upon receiving an uncompressed data, such as an image, a video, an audio, and/or a structured data, a machine learning model identifies an object in the uncompressed data such as a house, a dog, a text, a distinct audio signal, a unique data pattern, etc. The identified object is compressed using a compression treatment optimized for the identified object. The identified object, either before or after the compression, is removed from the uncompressed data. The uncompressed data with the identified object removed is compressed using a standard compression treatment.

Abstract: An imaging apparatus includes an imaging unit, a first image processor, and a record processor. The imaging unit sequentially acquires image data groups comprising pieces of image data different in exposure conditions to generate HDR moving image frames. The first image processor subjects the pieces of the image data belonging to each of the image data groups to first composition processing, to generate HDR moving image data comprising HDR moving image frames. The record processor is configured to record image data for still images so that the pieces of the image data belonging to each of the image data groups are subjected to second composition processing different from the first composition processing, to generate HDR still image data.

Abstract: Described is a system for identifying predictive keywords and generating a forecast. The system receives time-series of keyword counts (the time-series of keyword counts having a plurality of candidate keywords). The time-series of keyword counts are separated (i.e., marked or designated) into a group of time-series from active periods and a group of time-series from inactive periods. A covariance matrix is generated for each group of time-series. Generalized eigenvectors are generated between the two covariance matrices. Candidate keywords are ranked based on the generalized eigenvectors, such that candidate keywords having a rank exceeding a predetermined threshold are designated as predictive keywords. The predictive keywords are then provided to a machine learning system that generates a forecast based on the predictive keywords. Finally, a device is operated based on the forecast.

Abstract: The present disclosure relates to method and system for determining intent of subject using behavioural pattern by intent determination system. The intent determination system receiving video data associated with subject from one or more information sources, extracting pre-defined number of femes before target frame from video data of subject, obtaining plurality of behavioural parameters of subject for each of extracted frames, determining score for plurality of behavioural parameters of extracted femes using trained intent model, calculating weighted average score for plurality of behavioural parameters of extracted frames based on score of each of behavioural parameters, analysing emotion of subject for pre-defined number of frames based on weighted average score and determining intent of subject based on emotion of subject analysed. The present disclosure removes the need of human intervention in determining intent of subject.

Abstract: The present document is directed to methods and subsystems that identify and characterize document-containing subimages in a document-containing image. In one implementation, each type of document is modeled as a set of features that are extracted from a set of images known to contain the document. To locate and characterize a document subimage in an image, the currently described methods and subsystems extract features from the image and then match model features of each model in a set of models to the extracted features to select the model that best corresponds to the extracted features. Additional information contained in the selected model is then used to identify the location of the subimage corresponding to the document and to process the document subimage to correct for a variety of distortions and deficiencies in order to facilitate subsequent data extraction from the corrected document subimage.

Abstract: Video information defining video content may be accessed. The video content may include video frames. The video frames may be grouped into similarity groups based on similarity of the video frames. The similarity groups may include a first similarity group including a first video frame and a second video frame. The first video frame may be similar to the second video frame. Representative video frames may be selected from the similarity groups. An individual representative video frame may correspond to an individual similarity group. Representative video frames may include a first representative video frame selected from and corresponding to the first similarity group. The video content may be analyzed based on the representative video frames.

Abstract: Provided are an image processing device and an electronic endoscope system that can acquire a high-quality endoscope observation image by performing sufficient noise reduction processing even if an image capture environment changes. A noise reduction processing unit sets a pixel of interest and a neighboring pixel out of a plurality of pixels that constitute an endoscope observation image, and performs noise reduction processing on the pixel of interest according to a magnitude relationship between: a difference in pixel value between the pixel of interest and the neighboring pixel; and a reference threshold value. A reference threshold value changing processing unit changes the reference threshold value for noise reduction processing based on a ratio between at least two color components that are included in the pixel of interest.

Abstract: An intelligent assistant can be used to facilitate an end user's participation in a virtual meeting. The intelligent assistant can receive recognized speech from a speech recognition engine, and then forward the recognized speech to a natural language understanding engine which in turn ascribes a meaning to the recognized speech. The ascribed meaning can include a semantic representation of the recognized speech and thereby a semantic representation of voice commands enshrined within the recognized speech. Using the ascribed meaning, the intelligent assistant then searches for meeting information associated with the virtual meeting and uses that meeting information to join the end user to the virtual meeting. Meeting information can include a conference call number, a participant code, a moderator code or a link to an online meeting.

Abstract: An apparatus includes input(s) for obtaining a first image and for obtaining a plurality of image parameters that includes at least a first image parameter and a second image parameter; a simulated image creation module configured for computing a plurality of simulated images based on the first image and the plurality of image parameters, the plurality of simulated images having at least a first simulated image computed based on the first image parameter, and a second simulated image computed based on the second image parameter, wherein the first simulated image has a first image quality, the second simulated image has a second image quality, and the first image quality is different from the second image quality; an image registration module configured for performing a plurality of image registrations using the simulated images; and a non-transitory medium configured for storing results from the act of performing the plurality of image registrations.

Abstract: A method of converting a vector corresponding to an input image includes receiving input vector data associated with an input image including an object; and converting the received input vector data into feature data based on a projection matrix having a fixed rank, wherein a first dimension of the input vector data is higher than a second dimension of the feature data.

Abstract: In some embodiments of the present disclosure, a system for processing three-dimensional face scan data is provided. A three-dimensional scanner produces an image of a face including an area of interest that includes an eyebag area. A profile of the eyebag area is determined by the system. In some embodiments, the profile is determined based on a vertical slice at the center of the eyebag area. Profiles for multiple sets of scan data may be compared to determine quantitative differences between eyebag profiles. These differences may be used for quantitatively comparing the effects of products applied to the eyebag area between scans. These differences may also be used for predictively generating three-dimensional models to illustrate predicted effects of the use of a product on a face.

Abstract: A sample counter in each channel performs counting operation at a given rate. Independently for each channel, the rate and an initial value for the counter are set, and start and stop of the counting operation of the counter are controlled, so that a partial portion of an original waveform corresponding to a count range from the set initial value to a count stop point is reproduced in the channel. A control section sets the initial values in individual ones of a set of channels, selected from among the channels, such that sample values at different sample positions of the original waveform are simultaneously retrieved in individual ones of the set of channels, and controls an overlap adder to add up the retrieved sample values, so that sample values of an audio waveform signal with a plurality of partial portions of the original waveform, partially overlapping each other are output.

Abstract: An image processing apparatus includes an obtainer and an acquirer. The obtainer obtains video information including video of a measurement target and a visible light signal that includes a measurement time and measurement information of the measurement target measured by a measuring apparatus attached to the measurement target, the visible light signal being transmitted by the measuring apparatus. The acquirer acquires the measurement time and the measurement information from the visible light signal included in the video information obtained by the obtainer, and synchronizes, with respect to time, the measurement information and the video information based on the measurement time acquired.

Abstract: Example embodiments include a method of reducing noise include forming a main signal and one or more reference signals at a beam-former based on at least two received audio signals, detecting voice activity at a voice activity detector, where the voice activity detector receives the main and reference signals and outputting a desired voice activity signal, adaptively cancelling noise at an adaptive noise canceller, where the adaptive noise canceller receives the main, reference, and desired voice activity signals and outputs an adaptive noise cancellation signal, and reducing noise at a noise reducer receiving the desired voice activity and adaptive noise cancellation signals and outputting a desired speech signal.

Abstract: The present disclosure relates to a high precision method, model, and apparatus for calibrating the intrinsic camera parameter values of one or more intrinsic parameters of a camera. The intrinsic parameters include focal lengths, skew, optical center, and radial distortion coefficients. The practical implementation of the solution consists of a camera capturing several images of a planar calibration pattern from different viewpoints, and a computer implemented processing pipeline. The solution does not require additional human input and is robust to misdetection of the calibration pattern in some of the images.

Abstract: A system configured to improve human presence detection and/or localization by generating aggregate confidence values. The system may aggregate confidence values corresponding to overlapping regions of interest. The system may perform human presence detection by comparing the aggregate confidence values to a universal threshold, with aggregate confidence values above the universal threshold indicating that human presence is detected. The system may use the aggregate confidence values to generate a heatmap, may identify a strongest cluster of pixels in the heatmap and determine a bounding box surrounding the strongest cluster. To distinguish the strongest cluster from a second strongest cluster of pixels, the system may apply a Gaussian kernel with varying threshold values. The system may store the threshold value at which the strongest cluster separates from the second strongest cluster and may use the threshold value to generate the bounding box.

Abstract: Methods, systems, and apparatus, including an apparatus for determining pixel coordinates for image transformation and memory addresses for storing the transformed image data. In some implementations, a system includes a processing unit configured to perform machine learning computations for images using a machine learning model and pixel values for the images, a storage medium configured to store the pixel values for the images, and a memory address computation unit that includes one or more hardware processors. The processor(s) are configured to receive image data for an image and determine that the dimensions of the image do not match the dimensions of the machine learning model. In response, the processor(s) determine pixel coordinates for a transformed version of the input image and, for each of the pixel coordinates, memory address(es), in the storage medium, for storing pixel value(s) that will be used to generate an input to the machine learning model.

Abstract: A system and method for representing quasi-periodic waveforms, for example, representing a plurality of limited decompositions of the quasi-periodic waveform. Each decomposition includes a first and second amplitude value and at least one time value. In some embodiments, each of the decompositions is phase adjusted such that the arithmetic sum of the plurality of limited decompositions reconstructs the quasi-periodic waveform. Data-structure attributes are created and used to reconstruct the quasi-periodic waveform. Features of the quasi-periodic wave are tracked using pattern-recognition techniques. The fundamental rate of the signal (e.g., heartbeat) can vary widely, for example by a factor of 2-3 or more from the lowest to highest frequency. To get quarter-phase representations of a component (e.g.

Abstract: A device and a method for editing text in a portable terminal are provided. The device includes displaying content comprising at least one unit of text, receiving an input of at least one unit of text through a first display area, changing to an editing mode for editing the content when the input text is released from the first display area while dragging the input text to replace a specific word included in the content located at a second display area, and replacing the specific word with the input text when the input text is completely dragged to the specific word.

Abstract: Techniques for customizing a display of content related to dictionary entries are described herein. In one example, a user selects a word from a rendered electronic book. In response, the device outputs a dictionary user interface that includes display elements associated with a particular user. Preferences of a user may be utilized to determine what content (e.g., what display elements) to display within the dictionary UI as well as how the content is to be displayed. The preferences may be user-specified and/or learned from user behavior. For instance, a machine learning mechanism may be used to determine what content is the most interesting to a user.