Patents Examined by Jonathan S Lee
  • Patent number: 10325412
    Abstract: Methods and devices for cutting a three-dimensional image are disclosed. According to one example of the methods, a region of interest may be selected from an original three-dimensional image. A mesh model of the region of interest may be constructed. Then, distance field information and a shape index characteristic value of each mesh point in the mesh model may be obtained. A set of initiation points and a set of termination points may be obtained by splitting the mesh model according to the distance field information. Afterward, a mixed cost value of each mesh point may be obtained according to the distance field information, the shape index characteristic value and an image value of each mesh point. In this way, a cutting path may be determined from the set of initiation points and the set of termination points according to the mixed cost value.
    Type: Grant
    Filed: November 3, 2016
    Date of Patent: June 18, 2019
    Assignee: Shenyang Neusoft Medical Systems Co., Ltd.
    Inventors: Bingsheng Li, Wei He
  • Patent number: 10319075
    Abstract: With known image upscaling methods of noisy images, important detail information is lost during denoising. A method for upscaling noisy input images comprises upscaling a noisy input image to obtain a noisy High-Resolution (HR) image, denoising the noisy input image to obtained a denoised Low-Resolution (LR) image, upscaling the denoised LR image to obtain an upscaled denoised LR image, and combining the noisy HR image and the upscaled denoised LR image to generate a denoised HR image.
    Type: Grant
    Filed: November 4, 2016
    Date of Patent: June 11, 2019
    Assignee: InterDigital CE Patent Holdings
    Inventors: Mangesh Kothule, Jordi Salvador Marcos
  • Patent number: 10311628
    Abstract: One embodiment of the present invention includes a method for rendering a geometry object in a computer-generated scene. A screen space associated with a display screen is divided into a set of regions. For each region; a first sampling factor in a horizontal dimension is computed that represents a horizontal sampling factor for pixels located in the region, a second sampling factor in a vertical dimension is computed that represents a vertical sampling factor for the pixels located in the region, a first offset in the horizontal dimension is computed that represents a horizontal position associated with the region, and a second offset in the vertical dimension is computed that represent a vertical position associated with the region. When the geometry object is determined to intersect more than one region, an instance of the geometry object is generated each region that the geometry object intersects.
    Type: Grant
    Filed: July 30, 2018
    Date of Patent: June 4, 2019
    Assignee: NVIDIA CORPORATION
    Inventors: Eric B. Lum, Justin Cobb
  • Patent number: 10303982
    Abstract: In various embodiments, training objects are classified by human annotators, psychometric data characterizing the annotation of the training objects is acquired, a human-weighted loss function based at least in part on the classification data and the psychometric data is computationally derived, and one or more features of a query object are computationally classified based at least in part on the human-weighted loss function.
    Type: Grant
    Filed: July 25, 2018
    Date of Patent: May 28, 2019
    Assignee: PRESIDENT AND FELLOWS OF HARVARD COLLEGE
    Inventors: David Cox, Walter Scheirer, Samuel Anthony, Ken Nakayama
  • Patent number: 10304169
    Abstract: Provided is a method and device for correction restoration and analysis alarming of a distorted image. The method includes that: an original distorted image acquired by a distortion lens is received, and original distorted coordinates of each coordinate point in the original distorted image are acquired; and pre-stored distortion parameters of the distortion lens are acquired, and restored coordinates of each coordinate point in the original distorted image are determined to obtain a restored image.
    Type: Grant
    Filed: April 15, 2014
    Date of Patent: May 28, 2019
    Assignee: ZTE CORPORATION
    Inventor: Bin Chen
  • Patent number: 10296815
    Abstract: A convolutional neural network system for detecting at least one object in at least one image. The system includes a plurality of object detectors, corresponding to a predetermined image window size in the at least one image. Each object detector is associated with a respective down-sampling ratio with respect to the at least one image. Each object detector includes a respective convolutional neural network and an object classifier coupled with the convolutional neural network. The respective convolutional neural network includes a plurality of convolution layers. The object classifier classifies objects in the image according to the results from the convolutional neural network. Object detectors associated with the same respective down-sampling ratio define at least one group of object detectors. Object detectors in a group of object detectors being associated with common convolution layers.
    Type: Grant
    Filed: April 20, 2017
    Date of Patent: May 21, 2019
    Assignee: Ramot at Tel Aviv University Ltd.
    Inventors: Lior Wolf, Assaf Mushinsky
  • Patent number: 10289951
    Abstract: Methods and systems are provided for deblurring images. A neural network is trained where the training includes selecting a central training image from a sequence of blurred images. An earlier training image and a later training image are selected based on the earlier training image preceding the central training image in the sequence and the later training image following the central training image in the sequence and based on proximity of the images to the central training image in the sequence. A training output image is generated by the neural network from the central training image, the earlier training image, and the later training image. Similarity is evaluated between the training output image and a reference image. The neural network is modified based on the evaluated similarity. The trained neural network is used to generate a deblurred output image from a blurry input image.
    Type: Grant
    Filed: November 2, 2016
    Date of Patent: May 14, 2019
    Assignee: Adobe Inc.
    Inventors: Oliver Wang, Jue Wang, Shuochen Su
  • Patent number: 10269101
    Abstract: A restoration processing unit performs a restoration process based on an optical transfer function of an optical system for a target image. The entire angle of view of the optical system is greater than 90 degrees. A light amount evaluation region of the optical system is a region in which a distance from the center of an image formation plane is not less than 80% of half of the length of a diagonal line of an imaging surface of an imaging element. When a first evaluation wavelength is used, the ratio of the amount of light in the light amount evaluation region to the amount of light in a region of the optical system corresponding to the center of the image formation plane is not less than 25%. When a second evaluation wavelength is used, the value of the MTF of the optical system acquired at half of the Nyquist frequency of the imaging element is not less than 15%.
    Type: Grant
    Filed: October 13, 2017
    Date of Patent: April 23, 2019
    Assignee: FUJIFILM Corporation
    Inventors: Takuya Tanaka, Yousuke Naruse
  • Patent number: 10269098
    Abstract: According to embodiments of the invention, systems and methods are provided for computer-implemented removal of haze, fog, cloudiness and/or other undesired noise from digital imagery. The disclosed technology may be carried out automatically or semi-autonomously, with input from a user. Systems and methods generally involve detecting and isolating certain portions, regions, and/or objects found in a digital image. Depending on their detected depth of field and/or other parameters, certain changes are selectively implemented to reduce the appearance of the detected haze, fog, cloudiness and/or other undesired noise. User input may be employed to further improve certain characteristics of the digital image.
    Type: Grant
    Filed: November 1, 2016
    Date of Patent: April 23, 2019
    Inventor: Chun Ming Tsang
  • Patent number: 10261345
    Abstract: An imaging analysis apparatus includes: an information processing module, used to receive position and depth information of an imaging lens group and/or imaging receiver relative to an object, obtain an imaging parameter corresponding to the imaging lens group according to the position and depth information, and send the obtained imaging parameter. An imaging lens group, used to image an object, can include subregions having adjustable imaging characteristics; and a lens adjustment module, used to receive the imaging parameter of the imaging lens group, determine a subregion corresponding to the imaging parameter, and adjust the imaging characteristic of the subregion. An object can be imaged by using an imaging lens in which each subregion has adjustable imaging parameters, thereby adjusting and correcting a perspective deformation that occurs on the object, and preventing a perspective deformation from occurring on an image of the object acquired by a user.
    Type: Grant
    Filed: December 4, 2013
    Date of Patent: April 16, 2019
    Assignee: BEIJING ZHIGU RUI TUO TECH CO., LTD
    Inventors: Lin Du, Hongjiang Zhang
  • Patent number: 10255683
    Abstract: Techniques are generally described for discontinuity detection. A first and second frame of video data representing a portion of a physical environment are received. The first and second frames are divided into first and second blocks of pixels, respectively. A set of matching blocks are determined by matching each block of the first blocks with a corresponding block of the second blocks. A difference value is determined for each pair of matching blocks of the set of matching blocks. A mean difference value for the set of matching blocks. The mean difference value is convolved with a convolution kernel. A determination is made that a result of the convolving the mean difference value with the convolution kernel exceeds a mean threshold value. A determination is made that a discontinuity between the first frame and the second frame is not due to motion in the physical environment.
    Type: Grant
    Filed: October 31, 2017
    Date of Patent: April 9, 2019
    Assignee: AMAZON TECHNOLOGIES, INC.
    Inventors: Sharath Kumar Gopal, Rohith Mysore Vijaya Kumar
  • Patent number: 10255698
    Abstract: A system and method for CT image reconstruction are provided. The method may include: obtaining raw data set related to an object; generating a first image set based on the raw data set, wherein the first image set includes a first full quality image and a first max field of view image; generating one or more reference images based on the first max field of view image; generating a first bone information image based on the one or more reference images; generating a second image set based on the raw data set, wherein the second image set includes a second full quality image; generating a second bone information image based on the one or more reference images; correcting hardening beam artifact of the second full quality image based on the second bone information image to generate a hardening beam artifact corrected image.
    Type: Grant
    Filed: October 30, 2017
    Date of Patent: April 9, 2019
    Assignee: SHENZHEN UNITED IMAGING HEALTHCARE CO., LTD.
    Inventors: Hailing Peng, Dajun Wang, Weiqiang Guo
  • Patent number: 10235764
    Abstract: A method for collision detection between colliders in a real-time virtual scene includes: obtaining a shortest distance from a cylindrical collider corresponding to a virtual object to each plane of a first binary space partitioning collider in a virtual scene; extending each plane of the first binary space partitioning collider according to the shortest distance to form a second binary space partitioning collider; extending an axis-aligned bounding box of the first binary space partitioning collider to generate an outer edge-extended bounding box; obtaining an intersection set of the generated outer edge-extended bounding box and the second binary space partitioning collider; and performing collision detection between the cylindrical collider and a result of obtaining the intersection set of the outer edge-extended bounding box and the second binary space partitioning collider.
    Type: Grant
    Filed: October 30, 2017
    Date of Patent: March 19, 2019
    Assignee: TENCENT TECHNOLOGY (SHENZHEN) COMPANY LIMITED
    Inventor: Xini Kuang
  • Patent number: 10229501
    Abstract: A mobile robot and method for controlling the same are provided creating patches in images captured by a camera while the mobile robot is moving, estimating motion blur of the patches, and correcting the position of the mobile robot based on the patch from which the motion blur is eliminated, thereby increasing precision in tracking and reliability through accurate mapping. The mobile robot includes a main body, a traveler to move the main body, a camera combined with the main body to capture an image of a surrounding of the main body, a position detector to create a patch in the image captured by the camera, estimate a motion blur of the patch, and track a position of the main body based on the created patch from which the motion blur is eliminated, and a controller to control the traveler based on the position of the main body.
    Type: Grant
    Filed: November 1, 2016
    Date of Patent: March 12, 2019
    Assignee: SAMSUNG ELECTRONICS CO., LTD.
    Inventors: Bo Young Kim, No San Kwak, Kyung Shik Roh, Soon Yong Park, Suk June Yoon, So Hee Lee, Min Yong Choi
  • Patent number: 10223584
    Abstract: Video frames of a document are captured. A still image of the document is optimally automatically taken after the video frames have been recorded. The best image from the frames and the still image is determined. The best image is provided for OCR processing to derive document information represented in the best image for the document.
    Type: Grant
    Filed: November 1, 2016
    Date of Patent: March 5, 2019
    Assignee: NCR Corporation
    Inventor: Olufisayo Adebanke Ekpenyong
  • Patent number: 10204299
    Abstract: A computer-implemented method for training a deep learning network is presented. The method includes receiving a first image and a second image, mining exemplar thin-plate spline (TPS) to determine transformations for generating point correspondences between the first and second images, using artificial point correspondences to train the deep neural network, learning and using the TPS transformation output through a spatial transformer, and applying heuristics for selecting an acceptable set of images to match for accurate reconstruction. The deep learning network learns to warp points in the first image to points in the second image.
    Type: Grant
    Filed: November 3, 2016
    Date of Patent: February 12, 2019
    Assignee: NEC Corporation
    Inventors: Manmohan Chandraker, Angjoo Kim
  • Patent number: 10204427
    Abstract: A system and method for air calibration in a Computed Tomography (CT) imaging system are provided. A first set of data associated with air in a scanning area may be obtained. A second set of data associated with an object in the scanning area may be obtained. The second set of data based on the first set of data may be calibrated, and a set of reference values generated by a neural network model may be used to perform the calibration. A third set of data based on the calibration of the second set of data may be generated. Based on the third set of data, a CT image of the object may be generated.
    Type: Grant
    Filed: October 23, 2017
    Date of Patent: February 12, 2019
    Assignee: SHENZHEN UNITED IMAGING HEALTHCARE CO., LTD.
    Inventor: Dier Zhang
  • Patent number: 10194063
    Abstract: A device for capturing at least three views of an object with a camera includes a housing having a cavity. A slot in the housing communicates with the cavity and is adapted to accept an object. A plurality of mirrors in the cavity reflects at least three views of the object towards the camera, allowing the camera to capture an image of the views. A system for capturing multiple views of an object includes a housing having a cavity. A slot in the housing communicates with the cavity and is adapted to receive an object. A camera, a light source, and a plurality of mirrors in the cavity reflect at least three views of the object towards the camera, allowing the camera to capture an image three views.
    Type: Grant
    Filed: April 11, 2016
    Date of Patent: January 29, 2019
    Assignee: Lockmasters Security Institute, Inc.
    Inventors: J. Clayton Miller, Justin D. Seacat
  • Patent number: 10186013
    Abstract: An image generating apparatus for generating an output image based on an input panorama image, includes a parameter input unit inputting an output range parameter and a correction parameter, the output range parameter designating an output range in the panorama image, the correction parameter designating a correction part to be corrected in the output image; and an image correction unit correcting the correction part designated by the correction parameter in the output image. Further, the image correction unit calculates a similarity between the correction part and peripheral pixels and corrects the correction part based on the similarity and the peripheral pixels, and the output image is generated from the output range of the panorama image and is the panorama image corrected by the correction.
    Type: Grant
    Filed: December 11, 2014
    Date of Patent: January 22, 2019
    Assignees: Ricoh Company, Ltd., National University Corporation Nara Institute of Science and Technology
    Inventors: Daisuke Nakagawa, Norihiko Kawai, Tomokazu Sato, Naokazu Yokoya
  • Patent number: 10175147
    Abstract: Method and apparatus for checking tires for vehicle wheels. A tire to be checked is fed to a checking station. A sidewall of the tire lies on a supporting portion of a rotating table. The supporting portion lies on a plane while the rotating table has a rotation axis perpendicular to the plane. A centring operation is carried out to align, in the plane, the rotation axis of the tire with the rotation axis of the rotating table.
    Type: Grant
    Filed: November 29, 2016
    Date of Patent: January 8, 2019
    Assignee: PIRELLI TYRE S.P.A.
    Inventors: Vincenzo Boffa, Luca Tersi, Valeriano Ballardini