Abstract: According to one embodiment, a distance estimation device comprises a first distance estimation unit based on a first estimation method, and a second distance estimation unit based on a second estimation method different from the first estimation method. The second distance estimation unit is configured to change a part of the second estimation method according to an output of the first distance estimation unit.

Abstract: Examples disclosed herein may involve a computing system that is operable to (i) derive a first representation of an agent's trajectory from a first set of sensor data captured by a first sensor system associated with a vehicle, (ii) derive a second representation of the agent's trajectory from a second set of sensor data captured by a second sensor system associated with the vehicle, (iii) align the spatial reference frames of the first and second representations by applying a spatial transformation to a given one of the first and second representations, and (iv) align the temporal reference frames of the first and second representations by determining an origin-time offset between the temporal reference frames of the first and second representations and applying the determined origin-time offset to timing information encoded in the given one of the first and second representations.

Abstract: The invention relates to automatic discovery and evaluation of a motion by treating motion recognition as an optimization problem by considering a series of basic motions obtained by segmenting a subject motion. The method comprises segmenting time series data defining a motion of a subject into a plurality of segments, classifying each segment into a class for a basic motion by using time series data of the segment, and converting the motion of the subject to a sequence of high rank symbols in which each high rank symbol is formed from a series of the basic motions, wherein a function that calculates a score based on a set of a high rank symbol and a sequence of basic motions is provided and the motion of the subject is converted to the sequence of the high rank symbols by an optimization calculation using dynamic programming.

Abstract: Methods for programming memory cells of a resistive memory device include applying a voltage pulse sequence to a memory cell to set a logic state of the memory cell. An initial set sequence of voltage pulses may be applied to the memory cell, followed by a reform voltage pulse having an amplitude greater than the amplitudes of the initial set sequence, and within ±5% of the amplitude of a voltage pulse used in an initial forming process. Additional voltage pulses having amplitudes that are less than the amplitude of the reform voltage pulse may be subsequently applied. By applying a reform voltage pulse in the middle of, or at the end of, a memory set sequence including multiple voltage pulses, a resistive memory device may have a larger memory window and improved data retention relative to resistive memory devices programmed using conventional programming methods.

Abstract: The present disclosure provides an adaptive atmospheric correction method of hyperspectral satellite based on prior ground object spectral matching. The method includes: establishing an adaptive atmospheric parameter retrieval algorithm of hyperspectral satellite data; based on the estimation algorithm of the integrated atmospheric column water vapor content, using a function between a linear regression band ratio and a water vapor content for estimation; in the aerosol optical depth estimation method of the hyperspectral data, using the rich spectral information of the hyperspectral data to perform AOD iterative estimation with the ground object spectral library as prior knowledge in combination with spectral matching, and establishing an atmospheric correction algorithm combining with an MODTRAN radiative transfer model.

Abstract: A method and system for profiling a reference image and an object-of-interest therewithin is disclosed. The method includes storing semantic metadata that provides semantic information about environments for respective areas covered by respective Fields Of View (FOVs) of video cameras. The method also includes comparing the semantic metadata to additional metadata corresponding to at least one of an object accessory and a non-permanent object facet to identify a subset of the video cameras, less than an entire video camera universe, that meet a similarity score condition.

Abstract: A memory system includes: a semiconductor storage device including a memory cell array that includes memory cells and a temperature counter configured to increase a count value thereof at a rate that depends on a temperature of the memory cell array; and a memory controller configured to acquire the count value from the semiconductor storage device and reserve a refresh operation for a written memory cell of the memory cell array when a cumulative value of the count value, which is accumulated from when data was written to the memory cell to when the count value is acquired, exceeds a predetermined value.

Abstract: A method for whole-process trace leaving of video manuscript gathering, editing, and check, including: extracting a key frame from video clipping information, generating histogram data by using the key frame, and obtaining video modification information based on the histogram data; and on-chaining, by using timestamps as blockchain nodes, the video modification information completed at different time points to form a trace chain arranged according to the timestamp, to implement whole-process trace leaving for video clipping. The technical solutions of the present invention resolve a difficulty in whole-process trace leaving for video manuscript editing and issuing, and in addition, by using the blockchain techniques, implement tamper-proofing and security of trace information, thereby effectively improving security of gathering and editing in the media industry.

Abstract: A system includes a memory device including an block and a processing device, operatively coupled with the memory device, to perform operations including initiating a page scan with respect to a page of the block, determining whether to perform an erased page check, and in response to determining that the erased page check is not to be performed, performing a two-sided page scan with calibration feedback.

Abstract: A method for generating cropped images depicting objects captured in video data includes receiving, at a processor of a video camera system, a video stream including a series of video frames depicting an object. A classification for the object is generated, and an occurrence of the object being detected is identified in an additional video frame(s) from the series of video frames. A motion associated with the object is calculated based on the classification and the additional video frame(s). At least one image that depicts the object and includes a cropped portion of a video frame from the series of video frames is generated, along with an associated set of at least one quality score, in response to calculating the motion. The method also includes causing transmission of the at least one image to at least one remote compute device based on the set of at least one quality score.

Abstract: A method for identifying a dry salt flat based on the sentinel-1 data is provided. The method includes: carrying out field survey and acquiring data; processing sentinel-1 data; determining backscattering values, spectrum characteristics and colors of different polarization; and determining a classification criterion and carrying out threshold classification.

Abstract: A method for performing object detection during autonomous driving includes: performing 3D object detection in a 3D object detection segment; uploading an output of multiple sensors in communication with the 3D object detection segment to multiple point clouds; transferring point cloud data from the multiple point clouds to a Region Proposal Network (RPN); independently performing 2D object detection in a 2D object detector in parallel with the 3D object detection in the 3D object detection segment; and taking a given input image and simultaneously learning box coordinates and class label probabilities in a 2D object detection network operating to treat object detection as a regression problem.

Abstract: Disclosed is a storage device, which includes a nonvolatile memory device including a first memory block connected with a plurality of first word lines, and a memory controller connected with the nonvolatile memory device through a plurality of data lines. The memory controller sends a first command to the nonvolatile memory device through the plurality of data lines during a first command input period, sends a parameter to the nonvolatile memory device through the plurality of data lines during an address input period, and sends a second command to the nonvolatile memory device through the plurality of data lines during a second command input period. The nonvolatile memory device applies a turn-on voltage to all the plurality of first word lines connected with the first memory block based on the parameter during a first time in response to the first command and the second command.

Abstract: A method, processor, and computer readable medium for generating blended intensity data for processing by a further processor, the method comprising the steps of generating a local contrast map based on image data, the local contrast map comprising a plurality of contrast values. An alpha mask is generated by associating an alpha mask value to each contrast value. First intensity data based on the image data using a first intensity calculation method and second intensity data based on the image data using a second intensity calculation method are also generated. The blended intensity data is generated by blending from at least one of the first intensity data and the second intensity data based on the alpha mask values; and outputted to the further processor.

Abstract: A method for constructing a convolutional neural network model based on farmland images is applied in an electronic device. The method includes following steps: obtaining a number of farmland images of at least one farmland; obtaining a plurality of standard segmentation farmland images corresponding to each of the farmland images; dividing the farmland images and the standard segmentation farmland images into a training image set and a test image set; taking the farmland images and the standard segmentation farmland images in the training image set as input of a convolutional neural network, and constructing a convolutional neural network model based on the farmland images; and verifying the convolutional neural network model by using the farmland images in the test image set and the standard segmentation farmland images, and optimizing a plurality of parameters of the convolutional neural network model based on the farmland images.

Abstract: The present invention discloses a disparity estimation optimization method based on upsampling and exact rematching, which conducts exact rematching within a small range in an optimized network, improves previous upsampling methods such as neighbor interpolation and bilinear interpolation for disparity maps or cost maps, and works out a propagation-based upsampling method by the way of network so that accurate disparity values can be better restored from disparity maps in the upsampling process.

Abstract: An additive manufacturing method includes depositing a first amount of metal powder onto a powder bed of a printing system, spreading the first amount of metal powder across the powder bed to form a first layer, and depositing a first amount of binder material on the first layer. The additive manufacturing method also includes exposing the first layer to a first lighting condition, imaging the first layer under the first lighting condition to generate a first image, analyzing the first image of the first layer, and determining whether to adjust at least one printing parameter based on the analyzing.

Abstract: A system is provided. The system includes a multiply-and-accumulate circuit and a local generator. The multiply-and-accumulate circuit is coupled to a memory array and generates a multiply-and-accumulate signal indicating a computational output of the memory array. The local generator is coupled to the memory array and generates at least one reference signal at a node in response to one of a plurality of global signals that are generated according to a number of the computational output. The local generator is further configured to generate an output signal according to the signal and a summation of the at least one reference signal at the node.

Abstract: A method for detecting chromosomal abnormalities comprising conversion of a chromosomal image into an ideogram which is compared to a control or standard ideogram of a chromosome; or converting an ideogram of a chromosome into a chromosomal image.

Abstract: The present disclosure relates to a system (100) for generating a diagnostic model. The system (100) includes a processor (108) configured to analyse a plurality of reference data sets. The reference data sets each include at least one image (230, 240). The analysis identities at least one feature in each image (230, 240). A metric is calculated in dependence on the at least one identified feature. Outcome data associated with at least some of the reference data sets is acquired. The diagnostic model is compiled in dependence on the at least one calculated metric and the associated outcome data. The present disclosure also relates to a method of generating a diagnostic model; and a non-transitory computer-readable medium.