Patents by Inventor Koubun Sakagami

Koubun Sakagami has filed for patents to protect the following inventions. This listing includes patent applications that are pending as well as patents that have already been granted by the United States Patent and Trademark Office (USPTO).

  • Publication number: 20040085878
    Abstract: A multi-level data processing method converts a binary data into a multi-level data having n bits per symbol, where n is an integer satisfying n≧2. A {(n−1)×m}-bit binary data is arranged in upper (n−1) bits of multi-level data of m symbols, where m is an integer satisfying m≧2, and a (m−k)-bit binary data is converted into m bits according to a predetermined conversion rule and arranged in a lower 1 bit of the multi-level data of m symbols, where k is an integer satisfying m>k≧1, so as to convert a (n×m−k)-bit binary data into 1 set of multi-level data made up of m symbols.
    Type: Application
    Filed: October 21, 2003
    Publication date: May 6, 2004
    Inventors: Koubun Sakagami, Akihiko Shimizu
  • Publication number: 20040009371
    Abstract: A recording medium with a data format is provided. The data format includes frames and an ECC block. Each of the frames includes a plurality of n-level data, n being an integer satisfying n≧3, a clock mark that is n-level data for sampling the n-level data, and a frame sync that is n-level data for indicating a break point of the n-level data of a predetermined amount. The ECC block is a data block for error correction using a product code. The ECC block is formed by binary data of J×K words, where J is a natural number indicating the number of valid binary data in each frame, and K is a natural number indicating the number of frames when the n-level data are converted into binary data.
    Type: Application
    Filed: April 11, 2003
    Publication date: January 15, 2004
    Inventors: Koubun Sakagami, Yuichi Kadokawa
  • Publication number: 20030169666
    Abstract: A method of generating an efficient sequence of “nm+m−1” level random numbers, each sub-sequence of “m” successive random numbers thereof having a different pattern, including the steps of generating a sequence of n-level random numbers based on an initial value, a number being generated at a time, determining whether a pattern of a sub-sequence is “used”, if the pattern of the sub-sequence is “used”, setting a different value to the initial value, if the pattern of the sub-sequence is not “used”, accepting the newly generated number and marking the pattern as “used”. All of the steps are repeated until all of “nm” patterns are exhausted. Since the newly generated number is rejected if the pattern of the sub-sequence is “used”, the sequence becomes the shortest one of which all of “nm” patterns are exhausted.
    Type: Application
    Filed: March 6, 2003
    Publication date: September 11, 2003
    Inventor: Koubun Sakagami
  • Publication number: 20030112667
    Abstract: A mark representing information of a multivalued level is recorded in each of cells arranged in a recording medium. Upon reproducing the multivalued level from a reproduction signal of the recording medium, the multivalued level of a present cell is judged by referring to a judged multivalued level of a preceding cell preceding the present cell or an ensuing cell ensuing the present cell, or by referring to the judged multivalued levels of both the preceding cell and the ensuing cell.
    Type: Application
    Filed: July 30, 2002
    Publication date: June 19, 2003
    Inventors: Akihiko Shimizu, Koubun Sakagami, Hiroyuki Miyata, Yasuhide Fujiwara
  • Publication number: 20030110444
    Abstract: In a data processing method of converting (n×m−1)-bit data to multi-level data of m symbols of n bits per symbol, n and m being integers satisfying n≧2 and m≧2, respectively, the m symbols each comprising n-bit data are arranged in m columns in an n×m matrix, with a value of each of the m symbols being set to an even or odd number, by creating the m symbols each including (n−1)-bit data of the (n×m−1)-bit data, arranging the m symbols in the m columns so that the {(n−1)×m}-bit data is arranged in upper-side {(n−1)×m} bits of the n×m matrix, and converting the remaining (m−1)-bit data of the (n×m−1)-bit data to m-bit data so that the m-bit data is arranged in a row in lower-side m bits of the n×m matrix.
    Type: Application
    Filed: September 20, 2002
    Publication date: June 12, 2003
    Inventors: Koubun Sakagami, Kazunori Takatsu, Akihiko Shimizu, Yuichi Kadokawa
  • Patent number: 5497194
    Abstract: An electronic camera includes an image converting device, a voice converting device, an image data compressing and expanding device, a voice data compressing and expanding device, a single digital/analog converter, a sample-holding device for sample-holding and outputting the analog sound signals, a pedestal level generating device for generating signals indicative of pedestal levels, a change-over switching device for connecting the voice data compressing and expanding device to the sample-holding means when causing the pedestal level generating device to output the signals indicative of pedestal levels and for connecting the image data compressing and expanding device to the single digital/analog converter when causing the pedestal level generating device to prevent the outputting of the signals indicative of the pedestal level.
    Type: Grant
    Filed: January 10, 1992
    Date of Patent: March 5, 1996
    Assignee: Ricoh Company, Ltd.
    Inventors: Koubun Sakagami, Eiichi Maeda, Masafumi Tanaka
  • Patent number: 5142380
    Abstract: An image data processing apparatus includes a converter for subjecting each block to a discrete cosine transform (DCT) to convert blocks into DCT coefficients, a first memory for temporarily storing the DCT coefficients, a second memory for storing threshold values of a quantization matrix which is used for quantizing the DCT coefficients, a quantizing part for quantizing the DCT coefficients, a differential part for obtaining a difference in D.C. components of quantized DCT coefficients related to present and previous blocks, a zero detector for detecting whether or not A.C. components of the quantized DCT coefficient are zero coefficients, a counter for counting a number of successive zero coefficients of the A.C. components, a coder for subjecting the D.C. and A.C. components to a Huffman coding to output a coded D.C. and A.C.
    Type: Grant
    Filed: October 18, 1990
    Date of Patent: August 25, 1992
    Assignee: Ricoh Company, Ltd.
    Inventors: Koubun Sakagami, Masafumi Tanaka, Eiichi Maeda
  • Patent number: 5140322
    Abstract: A coder unit includes a coding circuit for converting information supplied from an external unit into code data having a length corresponding to the information, a connecting circuit for successively connecting to each other the code data obtained by the coding circuit and for outputting connected data obained by the connecting of the code data, a determination circuit for determining whether or not a number of bits of the connected data output from the connecting circuit has reached a predetermined value, and an output circuit for outputting a bit string code included in the connected data output from the connecting means as parallel data when the determination means determines that the number of bits of the connected data has reached the predetermined value, the bit string code being formed of bits starting from a leading end bit of the connected data, a number of bits of the bit string code being equal to the predetermined value.
    Type: Grant
    Filed: August 26, 1991
    Date of Patent: August 18, 1992
    Assignee: Ricoh Company, Ltd.
    Inventor: Koubun Sakagami
  • Patent number: 4977417
    Abstract: An electrophotography system for use in laser printers, printing machines and copying machines. The system includes a rototable drum having a photosensitive member which can take a first absorption spectrum state and a second absorption spectrum state, the variation from the first absorption spectrum state to the second absorption spectrum state being made by illuminating it with light whose intensity is above a predetermined threshold and whose wavelength is in a first predetermined wavelength region. Also included is an optical system including a light-emitting device for illumination of the photosensitive member so as to form thereon an electrostatic latent image, the optical system being arranged to directly illuminate the photosensitive member with light from the light-emitting device and further to illuminate it with reflected light from an original picture.
    Type: Grant
    Filed: December 14, 1989
    Date of Patent: December 11, 1990
    Assignee: Victor Company of Japan, Ltd.
    Inventors: Itsuo Takanashi, Toshikatsu Ichitou, Hiroki Kitamura, Takashi Yamamura, Koubun Sakagami, Atsushi Nakano