会员体验
专利管家(专利管理)
工作空间(专利管理)
风险监控(情报监控)
数据分析(专利分析)
侵权分析(诉讼无效)
联系我们
交流群
官方交流:
QQ群: 891211   
微信请扫码    >>>
现在联系顾问~
热词
    • 43. 发明专利
    • EXPANSION/REDUCTION DISPLAY DEVICE AND METHOD
    • JPH10312187A
    • 1998-11-24
    • JP12050697
    • 1997-05-12
    • TOSHIBA CORP
    • NATORI NAOKIYURA KOJI
    • G09G5/00G06T3/40G09G5/36
    • PROBLEM TO BE SOLVED: To realize the operation changing a display in the expansive or reductive direction in real-time by user's simple input operation and to mitigate a limit in display magnification. SOLUTION: When the user performs the operation pushing a left or right button of a mouse 111 while pushing a shift key of a key board 112 in the state that the user operates the mouse 111, and moves a mouse cursor on a picture area of a display device 140 to a required reference position, whether a real-time display in the expansive or reductive direction is instructed is detected by an expansive/reductive direction detection device 121. An expansive/ reductive picture generation device 123 calculates a display subject frame and the display magnification based on a lapse time and the reference position in the period when the operation is performed, and reads in the display data within the display subject frame from a virtual picture storage 130, and expands or reduces them at the calculated display magnification, and executes repeatedly the operation displaying on the picture area.
    • 44. 发明专利
    • PATTERN RECOGNIZING DEVICE
    • JPH06180773A
    • 1994-06-28
    • JP15044393
    • 1993-06-22
    • TOSHIBA CORP
    • NISHIMURA KAZUONATORI NAOKI
    • G06K9/62G06K9/66G06T7/00G06F15/70
    • PURPOSE:To obtain a pattern recognizing device capable of recognizing even a little incomplete input pattern also and eliminating the necessity of learning many patterns. CONSTITUTION:The pattern recognizing device is provided with a pattern analyzing part 1 for analyzing the similarity of a learning pattern for an input pattern inputted to an input part, a hypothesis pattern preparing part 4 for generating a pattern considered close to the input pattern as a hypothesis, a comparing means 5 for comparing the hypothesis pattern prepared by the generating part 4 with the input pattern, an attribute describing area 2 for describing the compared result of the comparing part 5 and the analyzed result of the analyzing part 1, and an inference part 3 for referring to description relating to the input pattern described in the area 2, changing the operation of the analyzing part 1 based upon the referred result, starting the generating part 4, and executing control for preparing a final judgement result from the contents of a final attribute describing area and sending the result to another device.
    • 46. 发明专利
    • DEVICE AND METHOD FOR RECOGNIZING CHARACTER
    • JP2000331119A
    • 2000-11-30
    • JP14330699
    • 1999-05-24
    • TOSHIBA CORP
    • NATORI NAOKI
    • G06K9/44G06K9/46G06K9/62G06K9/68
    • PROBLEM TO BE SOLVED: To eliminate the need for repetitive calculation in matching without giving any influence to matched results without selecting any model character. SOLUTION: This character recognizing device is constituted in such a way that a feature point and segment extracting section 13 extracts feature points and segments through an input section 11 which inputs information from the outside and a line thinning section 12 and a labeling section 14 labels each segment with the ID of the field. Then a template generating section 15 generates templates at every field based on the labeled segments and stores the templates in a template storing section 151 and a stroke extracting section 16 labels each segment of an input character based on the information stored in the storing section 151 and extracts the segments having the same label as one stroke. In addition, a discriminating section 17 discriminates the extracted stroke for truth and falsehood based on the knowledge stored in a knowledge storing section 171 and an output section 18 outputs the information on the extracted stroke which is discriminated as 'truth'.
    • 48. 发明专利
    • DEVICE AND METHOD FOR PATTERN RECOGNITION
    • JPH1097590A
    • 1998-04-14
    • JP25054996
    • 1996-09-20
    • TOSHIBA CORP
    • NATORI NAOKIHONMA TORU
    • G06K9/46
    • PROBLEM TO BE SOLVED: To shorten the time needed for pattern recognition and reduce the capacity of a dictionary, and further to perform stable feature extraction and stroke integration. SOLUTION: A line thinning processing for an input character pattern P is performed by a line thinning processing part 1, feature points and segments are extracted by a feature extraction part 2 according to the line-thinned image T, and a stroke integration part 3 obtains stroke information I on the basis of the feature point information and segment information S. A dictionary matching part 4 extracts a stable stroke from the stroke information I according to dictionary information D containing information on stable strokes which are registered in the dictionary 41 and have small influence of character deformation, matches it against the dictionary information D, and outputs the result R. In case of a failure in dictionary matching, a feature point correction part 22 corrects the feature point information C on a top-down basis according to the dictionary information D and a stroke correction part 32 corrects the stroke information, so that those corrected pieces of information will be used for rematching at the dictionary matching part 4.
    • 49. 发明专利
    • PATTERN RECOGNIZING DEVICE
    • JPH08315075A
    • 1996-11-29
    • JP13744195
    • 1995-05-12
    • GIJUTSU KENKYU KUMIAI SHINJOHOTOSHIBA CORP
    • NATORI NAOKI
    • G06K9/66G06F15/18G06N3/00G06N3/04
    • PURPOSE: To attain more accurate pattern recognize by detecting the partial area of a character pattern, storing the detected area and acquiring a feedback rule for rerecognizing the character pattern when a recognized result obtained from an output part is not determinate, and when a succeeding recognized result is not determinate, executing rerecognition by using the feedback rule. CONSTITUTION: A character pattern or a picture pattern inputted to an input part 1 is learned and an unleaned character pattern or picture pattern is sorted by an intermediate part 2. The intermediate part 2 summarizes information to be outputted and an output part 3 outputs a final recognized result. When the outputted recognized result is not determinate, a feedback control part 41 in a feedback part 4 detects the partial area of the character pattern and stores the detected area in a feedback data base 42. A feedback rule corresponding to the partial area is acquired by storing the area, and when a succeeding recognized result is not determinate, rerecognition is executed by using the feedback rule. Thereby more accurate recognition can be attained.
    • 50. 发明专利
    • DEVICE AND METHOD FOR COLLATING WORD STRING
    • JP2001291060A
    • 2001-10-19
    • JP2000102370
    • 2000-04-04
    • TOSHIBA CORP
    • NATORI NAOKI
    • G06K9/62G06F17/21G06F40/00G06K9/72
    • PROBLEM TO BE SOLVED: To provide a word string collating device and method capable of realizing much more highly precise word string collation at the time of collating an inputted word string with each word in a word dictionary in a character recognition field. SOLUTION: In a word string collating method for collating an inputted word string with address data in an address dictionary at the time of extracting a word string from the result of character recognition corresponding to a word string such as address information constituted of plural words, the inputted word string is made to correspond to the address data in the address dictionary, and a positional relation between the inter-word distance of the inputted word string and the inter-address data distance in the address dictionary which are made to correspond to each other is obtained, and whether or not the corresponding relation between the inputted word string and the address data in the address dictionary which are made to correspond to each other is optimal is decided based on the obtained positional relation and the inter-word distance of the word string.