Abstract: Methods and systems are provided for generating mattes for input images. A neural network system can be trained where the training includes training a first neural network that generates mattes for input images where the input images are synthetic composite images. Such a neural network system can further be trained where the training includes training a second neural network that generates refined mattes from the mattes produced by the first neural network. Such a trained neural network system can be used to input an image and trimap pair for which the trained system will output a matte. Such a matte can be used to extract an object from the input image. Upon extracting the object, a user can manipulate the object, for example, to composite the object onto a new background.

Abstract: The present disclosure discloses a method and an apparatus for providing a media file. A media request is received. In response to the media request, a biometric feature of the user is obtained. The biometric feature is detected by a biometric sensing device associated with a playback device. The target media file is selected by circuitry of an apparatus from a plurality of media files identified utilizing the detected biometric feature. The target media file is provided to the playback device for playback. The present disclosure resolves a technical problem of relatively low efficiency when providing relevant media files to users.

Abstract: A method and apparatus for multi-level data protection in a storage system. The storage system comprises a first storage device, a second storage device, and a third storage device, the method comprising: establishing a synchronous replication session between the first storage device and the second storage device so as to synchronously replicate data in the first storage device to the second storage device; establishing an asynchronous replication session between the first storage device and the third storage device so as to asynchronously replicate the data in the first storage device to the third storage device; and recovering the asynchronous replication by establishing an asynchronous replication session between the second storage device and the third storage device in response to a failure of the first storage device.

Abstract: Systems and methods provide for capturing a plurality of segments of an audio stream and, for each segment of the plurality of segments of the audio stream: performing feature extraction on an audio signal of the segment using a feature extraction machine learning model that analyzes the audio signal to generate a feature vector for the segment and generating a prediction value for the segment for whether there is music in the segment using the extracted feature vector and a music detector machine learning model. The systems and methods further provide for generating a probability value that there is music in the audio stream based on the prediction value for each of the plurality of segments and causing the audio stream to be identified based on determining that the probability value that there is music in the audio stream meets a predetermined threshold.

Abstract: Embodiments of the present disclosure provide a method and device for data duplication cutover. The method includes initiating duplication of initial data from a source device to a destination device, wherein the initial data are data of a file system stored on the source device upon the initiating. The method further includes, in response to completing the duplication of the initial data, updating a session associated with the file system. Furthermore, the method includes, after updating the session, triggering the source device and the destination device into a data unavailable state to perform duplication of delta data from the source device to the destination device, wherein the delta data are data of the file system changed during duplication of the initial data. Utilization of the cutover mechanism proposed in the present disclosure enables effective reduction of data unavailable period of time.

Abstract: Systems and methods are disclosed for selecting target objects within digital images. In particular, in one or more embodiments, the disclosed systems and methods generate a trained neural network based on training digital images and training indicators. Moreover, one or more embodiments of the disclosed systems and methods utilize a trained neural network and iterative user indicators to select targeted objects in digital images. Specifically, the disclosed systems and methods can transform user indicators into distance maps that can be utilized in conjunction with color channels and a trained neural network to identify pixels that reflect the target object.

Abstract: Methods and systems are provided for generating mattes for input images. A neural network system can be trained where the training includes training a first neural network that generates mattes for input images where the input images are synthetic composite images. Such a neural network system can further be trained where the training includes training a second neural network that generates refined mattes from the mattes produced by the first neural network. Such a trained neural network system can be used to input an image and trimap pair for which the trained system will output a matte. Such a matte can be used to extract an object from the input image. Upon extracting the object, a user can manipulate the object, for example, to composite the object onto a new background.

Abstract: Computer systems, devices, and associated methods of content pattern based automatic document classification are disclosed herein. In one embodiment, a method includes receiving a document and a sequence of words corresponding to a document class having a class label from a network storage. The method also includes determining a longest common subsequence of words between the words in the document and the sequence of words and calculating a similarity percentage between the document and the sequence of words based on the determined longest common subsequence. When the calculated similarity percentage is above a threshold, the class label corresponding to the document class is automatically applied to the received document in the network storage.

Abstract: Systems and methods are disclosed for dynamically adjusting the backlight of a display during video playback. Given an input video stream and associated minimum, average, or maximum luminance values of the video frames in the video stream, values of a function of the frame min, mid, or max luminance values are filtered using a temporal filter to generate a filtered output value for each frame. The instantaneous dynamic range of a target display is determined based on the filtered output value and the minimum and maximum brightness values of the display. A backlight control level is computed based on the instantaneous dynamic range, and the input signal is tone mapped by a display management process to be displayed on the target display at the selected backlight level. The design of a temporal filter based on an exponential moving average filter and scene-change detection is presented.

Abstract: The present invention discloses an automatic paper feeding method for a paper shredder and a paper shredder, including a step of disposing paper, a step of pressing paper, a step of automatically feeding paper continuously, and a step of shredding paper. The paper sheets to be shredded entering the automatic paper feeding channel are superimposed partially; the paper disposing mechanism and the automatic paper feeding channel of the paper shredder are located at two sides of the paper rolling device respectively; the paper rolling device is located at least at the position above the lower end portion of the paper disposing area/paper sheets to be shredded. In the present invention, the paper sheets to be shredded conveyed to the automatic paper feeding channel are superimposed partially; during shredding, the shredding cutter shreds a plurality of paper sheets to be shredded at a time, with a higher feed speed and an increased shred speed. Besides, the automatic feed speed can be adjusted.

Abstract: The present disclosure provides a cobra type femoral stem prosthesis, that is a femoral stem prosthesis for congenital dysplasia of the hip (CDH) which includes a femoral stem cone, a femoral neck and a stem body which are connected sequentially; the stem body includes proximal segment of femoral stem, middle segment of femoral stem and distal segment of femoral stem, wherein the femoral stem cone is a cone shape; One end of the femoral neck is connected with the proximal segment of femoral stem and there is a deflection angle therebetween and the other end of the femoral neck is connected with the large diameter end of the femoral stem cone.

Abstract: The present disclosure provides an artificial alpha femoral stem prosthesis which includes a femoral stem cone, a femoral neck and a stem body connected integratedly, wherein a tool operating hole is arranged at upper end of the the femoral neck; the stem body comprises a proximal segment of the femoral stem, a middle segment of the femoral stem and a distal segment of the femoral stem, and there is a deflection angle between the femoral neck and the stem body.

Abstract: Methods and systems for enhancing motion estimation are disclosed for stereoscopic video sequences. The motion vector parameters and disparity values for consecutive frames may be calculated to estimate motion in a coherent manner between two channels, such as a left channel and a right channel in stereoscopic images. Occlusion handling is also applied to improve the estimation.

Abstract: Systems and methods are disclosed for selecting target objects within digital images. In particular, in one or more embodiments, the disclosed systems and methods generate a trained neural network based on training digital images and training indicators. Moreover, one or more embodiments of the disclosed systems and methods utilize a trained neural network and iterative user indicators to select targeted objects in digital images. Specifically, the disclosed systems and methods can transform user indicators into distance maps that can be utilized in conjunction with color channels and a trained neural network to identify pixels that reflect the target object.

Abstract: A mosquito trap includes a container having an open end and a composition positioned in the container. The composition comprises at least one active agent. The mosquito trap also includes an enclosure protruding above the open end of the container. The enclosure includes an opening to access an interior of the enclosure and the open end of the container, and at least one active agent.

Abstract: The present invention discloses a household vacuum sealer and a bag-cutting sealing method. In the household vacuum sealer, at least one cutter device is slidably mounted on the upper hood, a heat-seal bar and an evacuative groove are arranged at the front side of the upper portion of the lower hood, the upper hood or the lower hood is provided with a bag receiving chamber, a detachable automatic rollback shaft is rotatably mounted in the bag receiving chamber, the elastic rollback mechanism for automatically rolling the rollback shaft back and the elastic unidirectional locking mechanism for locking the automatic rollback shaft are mounted at the end portion of the automatic rollback shaft: the cutter device protrudes from a bottom surface of the upper hood, and the cutter moves at one side of the heat-seal bar or one side of the evacuative groove back and forth after the upper hood is locked onto the lower hood, for cutting off the bag pressed tightly between the upper hood and the lower hood.

Abstract: Systems and methods are disclosed for dynamically adjusting the backlight of a display during video playback. Given an input video stream and associated minimum, average, or maximum luminance values of the video frames in the video stream, values of a function of the frame min, mid, or max luminance values are filtered using a temporal filter to generate a filtered output value for each frame. The instantaneous dynamic range of a target display is determined based on the filtered output value and the minimum and maximum brightness values of the display. A backlight control level is computed based on the instantaneous dynamic range, and the input signal is tone mapped by a display management process to be displayed on the target display at the selected backlight level. The design of a temporal filter based on an exponential moving average filter and scene-change detection is presented.

Abstract: Systems, methods, devices, media, and computer readable instructions are described for local image tagging in a resource constrained environment. One embodiment involves processing image data using a deep convolutional neural network (DCNN) comprising at least a first subgraph and a second subgraph, the first subgraph comprising at least a first layer and a second layer, processing, the image data using at least the first layer of the first subgraph to generate first intermediate output data; processing, by the mobile device, the first intermediate output data using at least the second layer of the first subgraph to generate first subgraph output data, and in response to a determination that each layer reliant on the first intermediate data have completed processing, deleting the first intermediate data from the mobile device. Additional embodiments involve convolving entire pixel resolutions of the image data against kernels in different layers if the DCNN.

Abstract: A file disposition review system analyzes multiple productivity platforms to generate an upcoming disposition report identifying data files that have reached a disposition review age associated with a file management label. A cross-platform manual review (CPMR) portal enables concurrent examination of file property data for data files identified in the upcoming disposition report. The CPMR portal further enables the disposition manager to simultaneously examine file property data corresponding to the data files identified in the upcoming disposition report and to issue disposition instructions. Exemplary disposition instructions may cause the multiple productivity platforms to permanently delete the identified data files without their underlying file content ever being transferred and/or copied from the particular productivity platforms that individual ones of the identified data files correspond to.