Vector Quantization Patents (Class 382/253)
-
Patent number: 8682086Abstract: Systems and methods process an image having a plurality of pixels includes an image sensor to capture an image; a first-layer to encode local patches on an image region; and a second layer to jointly encode patches from the same image region.Type: GrantFiled: May 18, 2011Date of Patent: March 25, 2014Assignee: NEC Laboratories America, Inc.Inventors: Yuanqing Lin, Kai Yu
-
Patent number: 8682823Abstract: Methods, systems and computer program products for resolving multiple magnitudes assigned to a target vector are disclosed. A target vector that includes one or more target vector dimensions is received. One of the target vector dimensions is processed to determine a total number of magnitudes assigned to the processed target vector dimension. Also, a source vector that includes one or more source vector dimensions is received. The received source vector is processed to determine a total number of features associated with the source vector. When it is detected that the total number of magnitudes assigned to the processed target vector dimension exceeds one, one of the assigned magnitudes is selected based on one of the determined features associated with the source vector.Type: GrantFiled: April 13, 2007Date of Patent: March 25, 2014Assignee: A-Life Medical, LLCInventors: Daniel T. Heinze, Mark L. Morsch
-
Patent number: 8666108Abstract: Each bit of information that is impressed upon a video signal is impressed upon each pixel of a block of pixels of the video signal a video signal so as to reduce errors in transmission of the additional information. Prior to impression, the additional information may be further coded, e.g., using channel coding, interleaving, and differential encoding. A receiver performs complimentary processes to recover the additional data.Type: GrantFiled: January 15, 2003Date of Patent: March 4, 2014Assignee: Alcatel LucentInventors: Mohsen Sarraf, Mohammad Hossein Zarrabizadeh
-
Patent number: 8655091Abstract: Computer implemented methods for compressing 3D hyperspectral image data having a plurality of spatial pixels associated with a hyperspectral image, and a number of spectral dimensions associated with each spatial pixel, include receiving, using a processor, the 3D hyperspectral image data, a set of basis vectors associated therewith, and either a maximum error amount or a maximum data size. The methods also include partitioning the 3D hyperspectral image data into a plurality of 2D images, each associated with one of the number of spectral dimensions, and an associated one of the set of basis vectors. The methods additionally include ranking the set of basis vectors if not already ranked. The methods may further include iteratively applying lossy compression to the 2D images, in an order determined by the ranking. Other embodiments and features are also disclosed.Type: GrantFiled: February 24, 2012Date of Patent: February 18, 2014Assignee: Raytheon CompanyInventors: Ian S. Robinson, Bradley A. Flanders
-
Publication number: 20140037224Abstract: Quantization for oversampled signals with an error minimization searches based upon clusters of possible sampling vectors where the clusters have minimal correlation and thereby decrease reconstruction error as a function of oversampling (redundancy) ratio.Type: ApplicationFiled: October 11, 2013Publication date: February 6, 2014Applicant: Texas Instruments IncorporatedInventor: Mohamed Mansour
-
Patent number: 8638978Abstract: Methods for embedding digital watermarks in compressed video include perceptual adapting a digital watermark in predicted and non-predicted data based on block activity derived from the compressed video stream, embedding in predicted objects in a video stream having separately compressed video objects, and bit rate control of watermarked video.Type: GrantFiled: July 28, 2009Date of Patent: January 28, 2014Assignee: Digimarc CorporationInventors: Adnan M. Alattar, Eugene T. Lin, Mehmet U. Celik
-
Patent number: 8630509Abstract: An image represented by multiple nodes can be processed by determining whether labels can be propagated to a node from another node of the image. Conceptually, an image can be presented as a “structured grid” of multiple nodes (e.g., a structured grid of pixels of an image). In a “structured grid,” two or more nodes of the same level (e.g., nodes in the same gray level) can determine in parallel whether to propagate a label from one or more of its neighboring nodes that are labeled and propagate one or more labels accordingly. An image can be processed by iteratively repeating this process for nodes of successive levels. It will be appreciated that the disclosed techniques allow parallelism without requiring partitioning of an image or having to merge partitioned images. The disclosed techniques are especially suited for watershed algorithms.Type: GrantFiled: January 13, 2010Date of Patent: January 14, 2014Assignee: Samsung Electronics Co., Ltd.Inventors: Bor-Yiing Su, Tasneem G. Brutch
-
Patent number: 8619861Abstract: Architecture that employs texture sensitive temporal filtering to reuse motion estimation information in a realtime encoder. The temporal filter is applied for classified static areas. The architecture reuses the motion estimation results on motion vectors, cost estimates (e.g., sum of absolute difference (SAD)), and edge awareness texture information to apply the temporal filter on the current picture. Filtering can be applied at the pixel level, block level or macroblock level.Type: GrantFiled: February 26, 2008Date of Patent: December 31, 2013Assignee: Microsoft CorporationInventors: Peter X. Zuo, Bruce Lin, Bharath Siravara, Walid Ali
-
Patent number: 8593687Abstract: A printer includes a display, a storing module, a picture decoder, a selecting module, a recorder, and a printing module. The storing module is configured to store a plurality of pictures. The picture decoder is configured to read a selected picture from the storing module, decode the selected picture and display the selected picture on the display. The selecting module is configured to draw a frame to surround a selected part of the selected picture. The recorder is configured to record the position and the size of the selected part. The printing module is configured to print the selected part.Type: GrantFiled: April 28, 2011Date of Patent: November 26, 2013Assignee: Hon Hai Precision Industry Co., Ltd.Inventor: Teng-Yu Tsai
-
Patent number: 8594366Abstract: A method of providing CABAC compliant changes such as watermarks comprises accessing encoded data such as video which comprise at least two blocks; creating or accessing a list of changes to the encoded data that include a direct change to a block; determining motion character or motion vector differential of non-immediate block, non-immediate blocks being adjacent to an immediate block that are immediately adjacent to the block; determining change to the immediate block based on original motion character of the block and the non-immediate block and the motion character of the block that would result from the application of the change; storing the change to the list if the change does not cause a difference to the immediate block; and evaluating other potential changes if other potential changes are available, wherein the other potential changes are subjected to the same process steps as the direct change.Type: GrantFiled: August 17, 2009Date of Patent: November 26, 2013Assignee: Thomson LicensingInventors: Dekun Zou, Jeffrey Adam Bloom, Shan He
-
Patent number: 8582908Abstract: Quantization for oversampled signals with an error minimization searches based upon clusters of possible sampling vectors where the clusters have minimal correlation and thereby decrease reconstruction error as a function of oversampling (redundancy) ratio.Type: GrantFiled: August 1, 2008Date of Patent: November 12, 2013Assignee: Texas Instruments IncorporatedInventor: Mohamed Mansour
-
Patent number: 8583898Abstract: A system and method of compiling program code, wherein the program code includes an operation on an array of data elements stored in memory of a computer system. The program code is scanned for operations that are vectorizable. The vectorizable operations are examined to determine whether they should be executed at least in part in a vector atomic memory operation (AMO) functional unit attached to memory. If so, the compiled code includes vector AMO instructions.Type: GrantFiled: June 12, 2009Date of Patent: November 12, 2013Assignee: Cray Inc.Inventor: Terry D. Greyzck
-
Publication number: 20130294707Abstract: A method for processing images includes identifying empiric model elements (EMEs) in an original high resolution photo-realistic image, where each EME includes a straight central segment, a color profile, and a control area; and geometrically modeling the EMEs in vectorized forms to achieve a generally full visual quality for a representation of said image.Type: ApplicationFiled: July 7, 2011Publication date: November 7, 2013Applicant: YEDA RESEARCH AND DEVELOPMENT LTDInventors: Yosef Yomdin, Dvir Haviv
-
Patent number: 8559737Abstract: An image coding method includes: (i) determining, for each of one or more associated blocks, whether to add a motion vector of the associated block to a list, and (ii) adding the motion vector of the associated block to the list when determining that the motion vector of the associated block is to be added to the list; selecting, from the list, a motion vector which is to be merged to a current block; and (i) merging the selected motion vector to the current block, and (ii) coding the current block using the merged motion vector as a motion vector of the current block, and in the determining, it is determined that the motion vector of the associated block is not to be added to the list when an associated picture and an associated reference picture match temporally or when a current picture and a current reference picture match temporally.Type: GrantFiled: June 27, 2012Date of Patent: October 15, 2013Assignee: Panasonic CorporationInventors: Toshiyasu Sugio, Takahiro Nishi, Youji Shibahara, Kyoko Tanikawa, Hisao Sasai, Toru Matsunobu
-
Patent number: 8533162Abstract: An object detection method that is provided with a step for extracting a plurality of reference feature vectors related to a local area from an image representing an object, and extracting a plurality of query feature vectors related to the local area from a search query image; a step for matching each query feature vector against each reference feature vector, and calculating a similarity score having a value that is higher the closer the distance between both vectors, the larger the local area for which the query feature vector has been extracted, and the larger the local area for which a matching reference feature vector has been extracted; a step for determining a reference feature vector for which a similarity score is highest as the similar vector for each query feature vector; and a step for acquiring a final score by object associated with the similar vectors, and setting the object returning the highest score as the detection result; and wherein the score is calculated by dividing a sum of the similaType: GrantFiled: August 16, 2010Date of Patent: September 10, 2013Assignee: Osaka Prefecture University Public CorporationInventors: Martin Klinkigt, Koichi Kise, Heiko Maus, Andreas Dengel
-
Patent number: 8520951Abstract: The present invention enables mixture of a core line vectorization process and a outline vectorization process, and comprises: inputting an object image of a vectorization process; extracting a core line of the object image; computing an original line width for every pixel of the core line acquired by the extracting; judging whether every section is indicative of an equal-width line or indicative of a surface figure by using the line width value for every pixel of the core line acquired by the estimating, the every unit delimiting the core line acquired by the extracting between two points of intersections and endpoints; separating a surface figure from the object image on the basis of a judging result of the judging; and approximating functionally the core line judged to be the equal-width line at the judging and a outline of the surface figure separated at the separating, respectively.Type: GrantFiled: April 29, 2009Date of Patent: August 27, 2013Assignee: Canon Kabushiki KaishaInventor: Yuuichi Tsunematsu
-
Patent number: 8497787Abstract: Systems and methods for decoding data using a hybrid decoder are provided. A data signal that includes a codeword is received. A signal quality indicator for the data signal is computed. One of a plurality of decoders is selected based on the computed signal quality indicator. Each of the plurality of decoders is configured to decode information based on a different decoding technique. The codeword included in the data signal is decoded using the selected one of the plurality of decoders.Type: GrantFiled: November 7, 2011Date of Patent: July 30, 2013Assignee: Marvell World Trade Ltd.Inventor: Xueshi Yang
-
Patent number: 8489395Abstract: A method and an apparatus for generating a lattice vector quantizer codebook are disclosed. The method includes: storing an eigenvector set that includes amplitude vectors and/or length vectors, where the amplitude vectors and/or length vectors are different from each other and correspond to a root leader of a lattice vector quantizer; storing storage addresses of the amplitude vectors and length vectors, where the amplitude vectors and length vectors correspond to the root leader and are in the eigenvector set; and generating a lattice vector quantizer codebook according to the eigenvector set and the storage addresses.Type: GrantFiled: November 28, 2011Date of Patent: July 16, 2013Assignee: Huawei Technologies Co., Ltd.Inventors: Haiting Li, Deming Zhang
-
Patent number: 8488681Abstract: This invention makes it possible to reduce power consumption of an encoding device while more appropriately encoding a video image in the first frame. An image capturing apparatus includes a camera unit which photo-electrically converts object light and outputs a video signal, an encoding unit which encodes the video signal, a camera information acquisition unit which acquires information associated with the operation status of the camera unit, and a calculation unit which calculates, on the basis of the information acquired by the camera information acquisition unit, an initial parameter serving as an encoding parameter for an initial encoding operation of the encoding unit before the start of an encoding operation of the encoding unit.Type: GrantFiled: August 14, 2006Date of Patent: July 16, 2013Assignee: Canon Kabushiki KaishaInventor: Satoru Kobayashi
-
Patent number: 8483500Abstract: Apparatus and methods for coding images geometric vector quantization (GVQ) having an over-complete dictionary which produces a sparse vector of coefficients as it contains large runs of zeros. The sparse encoding is particularly well suited for use with run-length entropy coding techniques. Image blocks are sparse coded using GVQ, with the vector of coefficients converted to RUN-LENGTH symbols, and binarized into a set of binary symbols. At least a portion of the binary symbols are used as contexts which can be selected when performing binary arithmetic coding of the binary coded RUN and LENGTH data to generate a bit stream containing the encoded image that provides enhanced compression.Type: GrantFiled: December 7, 2010Date of Patent: July 9, 2013Assignee: Sony CorporationInventors: Thai-Ha Nguyen, Ali Tabatabai
-
Patent number: 8477050Abstract: A system and method for redundant transmission is provided. In one embodiment, an input signal S is encoded as a list of fragments. Each fragment includes an index value and a projection value. The index points to an entry in a dictionary of signal elements. A repetition factor is assigned to each fragment based on its importance. After a fragment is added, a reconstructed signal is generated by decoding the list of fragments. Encoding terminates once the reconstructed signal is sufficiently close to the original signal S.Type: GrantFiled: September 15, 2011Date of Patent: July 2, 2013Assignee: Google Inc.Inventor: Pascal Massimino
-
Patent number: 8462170Abstract: The invention concerns image processing and, in particular, the processing of picture attribute fields for an image. A method of obtaining a new picture attribute field of an image is disclosed in which a picture attribute value at one position is allocated to a new position in the image in dependence upon the value of a parameter, such as luminance data, at the original position and at the new position and/or in dependence on the distance between the original position and the new position. The invention may be used to process picture attribute fields comprising: motion vectors; motion vector confidence; segment labels; depth labels; texture labels.Type: GrantFiled: July 20, 2007Date of Patent: June 11, 2013Assignee: Snell LimitedInventors: Jonathan Diggins, Michael James Knee
-
Patent number: 8447752Abstract: Sketch and tagging based image search may include receiving a sketch query input and identifying an object in a desired image. The object or objects of the sketch query may be tagged with a text, and searching performed based on the objects. Certain implementations include indexing patches of the images, where the patches represent the objects. Relevant images can be returned based on the index of the patches.Type: GrantFiled: September 16, 2010Date of Patent: May 21, 2013Assignee: Microsoft CorporationInventors: Changhu Wang, Zhiwei Li, Lei Zhang
-
Patent number: 8442118Abstract: The invention relates to a method for obtaining transformation parameters. A camera motion can be modeled by providing a vector field of motion vectors describing estimated motion vectors, projecting the vector field on at least one axis, and deriving the transformation vector parameters from the projection of the vector fields. As the camera motion can be modeled by way of translation, scale and rotation, the projections of the vector field on the axis can be used.Type: GrantFiled: May 19, 2006Date of Patent: May 14, 2013Assignee: Entropic Communications, Inc.Inventor: Wei Pien Lee
-
Patent number: 8437045Abstract: Provided are bitmap based trapping methods, apparatus and systems. According to one exemplary method, black trapping color image data is performed by estimating the continuous tone values associated with non-black pixels near a qualified black pixel and subsequently, the estimated continuous tone values are halftoned at the qualified black pixel locations and ORed with the original bitmap data.Type: GrantFiled: August 27, 2009Date of Patent: May 7, 2013Assignee: Xerox CorporationInventors: Meng Yao, John A. Handwork, Xing Li
-
Patent number: 8417034Abstract: Disclosed herein is an apparatus and method for separating a foreground and a background. The apparatus includes a background model creation unit for creating a code book including a plurality of code words in order to separate the foreground and the background, and a foreground/background separation unit for separating the foreground and the background using the created code book. The method includes the steps of creating a code book including a plurality of code words in order to separate the foreground and the background, rearranging the cord words of the created code book on the basis of the number of sample data that belong to each of the code words, and separating the foreground and the background using the code book.Type: GrantFiled: August 5, 2009Date of Patent: April 9, 2013Assignee: Electronics and Telecommunications Research InstituteInventors: Kap Kee Kim, Bon Woo Hwang, Ji Hyung Lee, Jin Seo Kim, Seung Wook Lee, Chang Woo Chu, Ho Won Kim, Bon Ki Koo, Gil haeng Lee
-
Patent number: 8401325Abstract: According to one aspect of the present invention, there is provided an image processing apparatus comprising: a thinning process unit which extracts a core line by applying a thinning process; a line width estimation unit which estimates an original line width of each pixel of the core line; a core line division unit which divides the core line; a monospaced line determination unit which determines, based on line width information, whether or not each of core lines divided by the core line division unit is a monospaced line; a connection relation information generation unit which generates connection relation information with another core line in association with each of the core lines; and a core line modification unit which modifies the core lines, based on pieces of line width information of core line pixels, a monospaced line determination result, and pieces of connection relation information generated.Type: GrantFiled: December 14, 2009Date of Patent: March 19, 2013Assignee: Canon Kabushiki KaishaInventor: Yuuichi Tsunematsu
-
Patent number: 8385670Abstract: The restoration of images by vector quantization utilizing visual patterns is disclosed. One disclosed embodiment comprises restoring detail in a transition region of an unrestored image, by first identifying the transition region and forming blurred visual pattern blocks. These blurred visual pattern blocks are compared to a pre-trained codebook, and a corresponding high-quality visual pattern blocks is obtained. The high-quality visual pattern block is then blended with the unrestored image to form a restored image.Type: GrantFiled: August 20, 2008Date of Patent: February 26, 2013Assignee: Microsoft CorporationInventors: Feng Wu, Xiaoyan Sun
-
Patent number: 8379725Abstract: A motion-vector searching method is provided. The motion-vector searching method includes the steps of: reading out observed processing blocks and reading out reference processing blocks; concurrently computing partial evaluation values for each of a plurality of position vectors; obtaining a total evaluation value for each particular one of the position vectors; and finding the motion vector as an optimum position vector determined by comparison of the total evaluation values each obtained for one of the position vectors with each other.Type: GrantFiled: April 11, 2007Date of Patent: February 19, 2013Assignee: Sony CorporationInventor: Naoki Kuzuya
-
Patent number: 8374452Abstract: The GCL sequence index of each cell/sector is assigned so that a difference of the indices between adjacent cells/sectors is more than N, where N?1. During GCL index detection, the adjacent N?1 indices of an index with the largest magnitude can be deleted from handoff candidate cell indices.Type: GrantFiled: April 7, 2011Date of Patent: February 12, 2013Assignee: Motorola Mobility LLCInventors: Hidenori Akita, Masaya Fukuta, Hiroshi Hayashi
-
Patent number: 8358692Abstract: In order to set a quantization step according to a target amount of code with high precision, a coefficient set according to the resolution of a moving picture to be encoded and the encoding method of a frame to be encoded is selected from a plurality of coefficient sets corresponding to the resolutions of moving pictures and encoding methods of frames. A feature amount of the frame to be encoded is extracted, and when the frame to be encoded undergoes inter-frame coding, a function indicating the relationship between a quantization step and amount of generated data is generated based on the selected coefficient set and feature amount. When the function is generated, a quantization step according to a target amount of code is set based on that function in encoding of the frame to be encoded.Type: GrantFiled: April 25, 2008Date of Patent: January 22, 2013Assignee: Canon Kabushiki KaishaInventor: Masaki Suzuki
-
Patent number: 8341671Abstract: Disclosed herein is a system and method for providing a supplementary information associated with a broadcast content. The system provides a supplementary information associated with a specific broadcast program in the broadcast content by means of a supplementary information server, wherein the supplementary information server comprises a broadcasting receiving unit for receiving the broadcast content; a synchronization information unit for storing a synchronization information of the supplementary information associated with the specific broadcast program; a video comparison unit for comparing screens of the broadcast content with a previously stored scene change screen of the specific broadcast program and setting a start time of the specific broadcast program; a supplementary information extracting unit for calculating an elapsed time from the start time of the specific broadcast program to a specific screen and extracting the supplementary information corresponding to the elapsed time.Type: GrantFiled: January 8, 2008Date of Patent: December 25, 2012Assignee: SK Telecom Co., Ltd.Inventors: Goon Seop Lee, Dong Hahk Lee, Jae Hwang Yu, Jong Tae Ihm, Se Hyun Oh, Kang Il Koh
-
Patent number: 8340439Abstract: In an image conversion method, a value which reflects the mutual relationship between the classes of pixel patterns each formed from a pixel classified as one of a plurality of classes and peripheral pixels is set as a converted value corresponding to each of the plurality of classes, a pixel of interest is sequentially selected from the input image, and a pixel pattern formed from the selected pixel of interest and a predetermined number of pixels around it is classified as one of the plurality of classes in accordance with a neighboring pattern obtained based on the relationship between the value of the pixel of interest and the values of peripheral pixels located at predetermined relative positions with respect to the pixel of interest. The value of the pixel of interest is converted into a converted value set for a class to which the pixel of interest has been classified.Type: GrantFiled: June 18, 2010Date of Patent: December 25, 2012Assignee: Canon Kabushiki KaishaInventors: Yusuke Mitarai, Masakazu Matsugu, Katsuhiko Mori
-
Patent number: 8340450Abstract: A vector quantizer includes a lattice quantizer (10) approximating a vector x by a lattice vector belonging to a lattice ?0. A lattice vector decomposer (14) connected to the lattice quantizer successively decomposes the lattice vector into a sequence of quotient vectors y, and a sequence of remainder vectors ri on successive lattices ?I?1 by lattice division with a corresponding predetermined sequence of integers pi?2, where i=1 . . . k and k is a positive integer representing the number of elements in each sequence.Type: GrantFiled: September 12, 2006Date of Patent: December 25, 2012Assignee: Telefonaktiebolaget LM Ericsson (Publ)Inventor: Anisse Taleb
-
Patent number: 8340451Abstract: Provided is a method for constructing an image database for object recognition, which includes a feature extraction step of extracting local descriptors from object images which are to be stored in an image database, a scalar quantization step of quantizing a numeric value indicating each dimension of each of the local descriptors into a predetermined number of bit digits, and a storing step of organizing each of the local descriptors after the quantization to be able to be searched for in the closest vicinity, giving to the local descriptor an identifier of the image from which the local descriptor has been extracted, and storing the local descriptor to which the identifiers are given in the image database.Type: GrantFiled: April 27, 2009Date of Patent: December 25, 2012Assignee: Osaka Prefecture University Public CorporationInventors: Kazuto Noguchi, Koichi Kise, Masakazu Iwamura
-
Patent number: 8335260Abstract: A method for quantizing vector. The method includes: performing a quantization process on a vector to be quantized by use of N basic codebook vectors and the adjustment vectors of each of the basic codebook vectors, generating a basic codebook vector and an adjustment vector used for quantizing the vectors to be quantized, N being a positive integer larger than or equal to 1. According to the present invention, based on the method a device for quantizing vector is disclosed. According to embodiments of the present invention, the quantization of an input vector is done by introducing the modification vectors for the base codebook vectors, therefore the memory amount of the base codebook vectors is reduced effectively, and the calculation amount is merely the calculation amount required for going through N codebooks. Therefore, the complexity of the vector quantization could be decreased effectively.Type: GrantFiled: November 18, 2008Date of Patent: December 18, 2012Assignee: Huawei Technologies Co., Ltd.Inventors: Lixiong Li, Libin Guo, Liang Zhang, Dejun Zhang, Wehai Wu, Tinghong Wang
-
Patent number: 8331708Abstract: Methods and apparatuses for compressing and decompressing digital data. The method for compressing digital data comprises a number of steps: a) generating a vector-valued dataset according to the digital data, b) transforming the vector-valued dataset into multiwavelet coefficients, and c) entropically coding the multiwavelet coefficients. The method for decompressing digital data is substantially made up of the same steps as the method for compressing digital data but functioning in a reverse manner.Type: GrantFiled: January 18, 2007Date of Patent: December 11, 2012Assignee: Numeri Ltd.Inventor: Ilan Bar-On
-
Patent number: 8331447Abstract: A decoding apparatus includes a standard predicted image generating unit which generates a standard decoded image and a standard predicted image of standard image quality; a non-standard decoded image generating unit which generates a non-standard decoded image different in image quality based on a prediction error information; a non-standard predicted image generating unit which generates a non-standard predicted image different in image quality based on the non-standard decoded image; and a correction value calculating unit which calculates correction values corresponding to differences between the standard and non-standard predicted images. The non-standard decoded image generating unit includes a predicted image reconstructing unit which corrects the non-standard predicted image, and a decoding unit which inverse-quantizes the prediction error information and adds the inverse-quantized prediction error information and the corrected non-standard predicted image to generate the non-standard decoded image.Type: GrantFiled: November 24, 2009Date of Patent: December 11, 2012Assignee: Oki Electric Industry Co., Ltd.Inventor: Satoshi Nakagawa
-
Patent number: 8320696Abstract: In order to reduce wasteful operation processes by operation units, a data supply unit reads out a plurality of pieces of pixel data from an image memory and temporarily stores the pixel data in registers, and outputs the pixel data while shifting the pixel data in one direction; product-sum operators concurrently perform operation processes based on multiple pieces of pixel data outputted from the data supply unit and filter coefficients in a filter kernel. At this time, a selector within the data supply unit selects the pixel data stored in each register in accordance with the type of the filter operation process, and outputs the selected pixel data to the product-sum operators.Type: GrantFiled: September 30, 2008Date of Patent: November 27, 2012Assignee: Canon Kabushiki KaishaInventors: Takahisa Yamamoto, Masami Kato
-
Publication number: 20120237134Abstract: Image descriptor quantization technique embodiments are presented which quantize an image descriptor defined by a vector of number elements. This is generally accomplished by lowering the number of bits per number element to a prescribed degree. The resulting quantized image descriptor exhibits minimal loss of matching reliability while at the same time reducing the amount of storage space needed to store the descriptor in a database. Lowering the number of bits per number element also allows for increased matching speed.Type: ApplicationFiled: June 2, 2012Publication date: September 20, 2012Applicant: MICROSOFT CORPORATIONInventor: Simon Winder
-
Patent number: 8270747Abstract: An image encoding device that receives a plurality of component values of a color input signal inputted to an object pixel, vector quantizes, and encodes each of the plurality of component values includes: a difference vector quantization unit that vector quantizes and encodes each difference value obtained by subtracting prediction component values based on a previous object pixel from the component values of the object pixel respectively; an absolute value vector quantization unit that vector quantizes and encodes each different value of the object pixel; a determination unit that compares each difference value with respective comparison ranges to determine a code selection; and a code selector that selects one of an output from the difference vector quantization unit and an output from the absolute value vector quantization unit based on an output from the determination unit.Type: GrantFiled: March 24, 2009Date of Patent: September 18, 2012Assignee: Seiko Epson CorporationInventor: Masakazu Isomura
-
Patent number: 8228983Abstract: The invention is used in video coding. Systems, apparatuses and methods for processing an order-16 integer transform from an order-8 transform are provided. The order-16 transform method involves expanding an order-8 transform by generating an order-16 integer matrix and a scaling matrix.Type: GrantFiled: December 4, 2007Date of Patent: July 24, 2012Assignees: Hong Kong Applied Science and Technology Research, Institute Company LimitedInventors: Wai-Kuen Cham, Chi Keung Fong, Jie Dong, King Ngi Ngan, Hoi Ming Wong, Lu Wang, Yan Huo, Thomas H. Y. Pun
-
Patent number: 8208731Abstract: Image descriptor quantization technique embodiments are presented which quantize an image descriptor defined by a vector of number elements. This is generally accomplished by lowering the number of bits per number element to a prescribed degree. The resulting quantized image descriptor exhibits minimal loss of matching reliability while at the same time reducing the amount of storage space needed to store the descriptor in a database. Lowering the number of bits per number element also allows for increased matching speed.Type: GrantFiled: April 7, 2008Date of Patent: June 26, 2012Assignee: Microsoft CorporationInventor: Simon Winder
-
Patent number: 8208718Abstract: In the present invention, processing for setting a parameter expressing a measurement condition of three-dimensional measurement to a value necessary to output a proper recognition result is easily performed. The three-dimensional measurement is performed to stereo images of real models WM1 and WM2 of a workpiece using a measurement parameter set by a user, and positions and attitudes of the workpiece models WM1 and WM2 are recognized based on the measurement result. An image expressing the recognition result is displayed, and numerical data indicating the selected recognition result is set to sample data in response to a user manipulation for selecting the recognition result.Type: GrantFiled: February 23, 2010Date of Patent: June 26, 2012Assignee: Omron CorporationInventors: Shiro Fujieda, Atsushi Taneno, Hiroshi Yano, Yasuyuki Ikeda
-
Publication number: 20120121194Abstract: A feature vector is encoded into a sparse binary vector. The feature vector is retrieved, for example from storage or a feature vector generator. The feature vector represents a media object or other data object. One or more permutations are generated, the dimensionality of the generated permutations equivalent to the dimensionality of the feature vector. The permutations may be generated randomly or formulaically. The feature vector is permuted with the one or more permutations, creating one or more permuted feature vectors. The permuted feature vectors are truncated according to a selected window size. The indexes representing the maximum values of the permuted feature vectors are identified and encoded using one-hot encoding, producing one or more sparse binary vectors. The sparse binary vectors may be concatenated into a single sparse binary vector and stored. The sparse binary vector may be used in the similarity search, indexing or categorization of media objects.Type: ApplicationFiled: November 3, 2011Publication date: May 17, 2012Applicant: GOOGLE INC.Inventor: Jay Yagnik
-
Patent number: 8179400Abstract: A method for controlling an ambient lighting element including receiving a content signal, analyzing the content signal to determine a motion vector of an object (120A) depicted in the content signal, presenting the content signal on a display device, and adjusting an ambient lighting effect provided by the ambient lighting element as determined by the motion vector The presented content signal may be portioned into macro-blocks (110A) and sub-blocks (230). A motion vector of each sub-block (230) may be resolved into components that are parallel and perpendicular to an outside edge of the display device. An average color of each sub-block (230) depicted within a macro-block (110A) in proximity to an outside edge of the display device may be weighted by the motion vector of the corresponding sub-block (230) for determining an average color of the macro-block (110A). The average color of the macro-block (110A) may be used to adjust the ambient lighting element.Type: GrantFiled: February 26, 2007Date of Patent: May 15, 2012Assignee: Koninklijke Philips Electronics N.V.Inventor: Cornelis Wilhelmus Kwisthout
-
Patent number: 8175149Abstract: Provided is a method and apparatus for controlling a bit-rates in which an amount of generated bits is adjusted according to the complexity of each frame, and a bit-rate within a frame can be further effectively adjusted by applying different quantization parameter (QP) values to respective macro blocks. In the apparatus for controlling a bit-rate, a motion compensated temporal filtering (MCTP) operation is performed for each group of pictures (GOP) in an input scalable layer, and thereafter bits are allocated to respective frames in consideration of frame type and frame complexity. Furthermore, a QP is adaptively determined in the unit of a macro block on the basis of the allocated bits.Type: GrantFiled: November 21, 2006Date of Patent: May 8, 2012Assignee: Electronics and Telecommunications Research InstituteInventors: Dong-San Jun, Jae Gon Kim, Jin Woo Hong, Jong-Ki Han, Nae-ri Park
-
Patent number: 8155445Abstract: The present invention relates to an image processing method, an image processing apparatus and an image processing program for dealing with inverted characters (outlined characters) constituted by white pixels on a black ground in a tree structure same as that of normal characters constituted by black pixels on a white ground. In the present invention, black pixel blocks and white pixel blocks are sampled recursively from a binary image, tree structure data indicating a positional relation between the sampled black pixel blocks and white pixel blocks is created, an inverted image is created by white-black-inverting the insides of black pixel blocks that can include inverted characters, of black pixel blocks included in the tree structure data, white pixel blocks and black pixel blacks are sampled from the created inverted image, and data regarding the sampled white pixel blocks and black pixel blocs is added to corresponding nodes of the tree structure data.Type: GrantFiled: September 25, 2007Date of Patent: April 10, 2012Assignee: Canon Kabushiki KaishaInventor: Tomotoshi Kanatsu
-
Patent number: 8150186Abstract: An amount-of-compressed-data control method applicable to image data compression processing for compressing an amount of data is disclosed. The method includes the steps of: performing quantization by: dividing one digital image into multiple blocks each having n×n pixels; performing orthogonal transform on each block; and dividing n×n conversion coefficients resulting from the conversion by each threshold value of a quantization matrix including n×n threshold values each resulting from the multiplication of a predetermined coefficient S (where S is a positive real number); and variable-length encoding the quantized data.Type: GrantFiled: November 28, 2007Date of Patent: April 3, 2012Assignee: Sony CorporationInventors: Masanori Imata, Toshihiko Hirose, Michio Hirai
-
Patent number: 8150177Abstract: An image processing apparatus which compresses an image, the apparatus includes: a storage unit which stores a code book, the code book being prepared by allocating identification codes to code blocks, the code blocks being formed by executing quantization processing regarding a plurality of first image blocks as multidimensional vectors, the plurality of first image blocks being generated by cutting, in preset sizes, first regions out of a code book preparation image containing a first edge, each of the first regions surrounding and including an edge point of the first edge, the first edge being detected from the code book preparation image; an edge detection unit which detects a second edge from a compression target image containing the second edge; an image block generation unit which cuts, in preset sizes, second regions out of the compression target image to generate a plurality of second image blocks, each of the second regions surrounding and including an edge point of the detected second edge; a searcType: GrantFiled: July 27, 2011Date of Patent: April 3, 2012Assignee: Kabushiki Kaisha ToshibaInventor: Tadashi Mitsui