Image analysis – Pattern recognition
Reexamination Certificate
2000-06-30
2002-05-07
Patel, Jayanti K. (Department: 2621)
Image analysis
Pattern recognition
C345S215000, C704S010000, C707S793000
Reexamination Certificate
active
06385339
ABSTRACT:
BACKGROUND OF THE INVENTION
The present invention relates to a pattern data processing system configured for various kinds of patterns, such as characters and voices, on the assumption of multi-user. More particularly, it concerns a pattern data processing system having a learning function.
Also, the present invention relates to a pattern recognizing method of and system for executing a recognizing process of entered character and figure patterns.
There was proposed a prior art of pattern recognizing system in, for example, the Japanese Patent Laid-Open No. 55-135973. The prior art is of learning method on the assumption of a particular, or a single, user. It executes learning of patterns given by the users in, for example, the way that necessary data are entered in a data base to process.
For the reason, the prior art has to be made to learn the patterns used in common by a plurality of users (members) of a group individually even if the common patterns are agreed by them. That is, if a mark or the like is to be used in common by the members of the group, to make it possible to enter it by hand-writing, the individual users who are the members of the group have to make it to be pattern-learnt.
As an example, the group uses a kana-kanji conversion of a word processor that is a tacit common agreement in the group. If a term “*BISHAMON” should be converted to “*BISHAMON,” for example, the term has to be entered individually. That is, to increase operationability of the own pattern data processing system, the user has to make by himself or herself, or ask another user to make the system learn the pattern. In other words, the operationability of the own pattern data processing system cannot be increased forever unless the user has to make, by himself or herself, or ask another user to make the system learn the pattern.
Therefore, to increase the operationability of the own pattern data processing system without making pattern learning by the user by himself or herself or asking another user to do it, it is a problem to allow another person of the group to use results of the pattern learning of the system.
It is preferable that the members of the group should learn tacit, or expressed, common agreements to omit their individual operations and increase the operationability. That is, a pattern data processing system for multi-user should be preferably made for generally wide use by not only a single user, but also a plurality of users.
For the purpose, there are problems: one is to extract patterns that would be used in common by a plurality of users of a group, and the other is to make the common patterns to be learnt as the ones belonging to the group.
Prior filing systems, such as the UNIX, are made to allow all the members of the group to read and write the files in the way that the files are prepared to have a group attribute. However, the prior filing system art alone has the disadvantage that a system manager of the group must make a plan in what a way and how data should be rewritten, or learnt by a dictionary, to increase the operationability.
There was proposed another prior art of pattern recognizing system in the Japanese Patent Laid-Open No. 56-99583. This prior art is a method of obtaining a recognition result with use of a plurality of recognizing arrangements for respective characters to recognize them accurately depending on pattern. The method obtains candidates based on appearance rates. However, the method has the disadvantage that if the appearance rate is low, recognition cannot be made even at what a high accuracy a pattern is entered.
There was proposed still another prior art of pattern recognizing system in the Japanese Patent Publication No. 62-15911. This prior art is a method that a dictionary has description of what parts form what shape is. There was proposed still another prior art of pattern recognizing system in the Japanese Patent Laid Open No. 2-56689. This prior art discloses method is which a dictionary includes a description of what stroke are continuating of what shapes. These prior arts are based on that a recognition result can be made close to that of man or woman in the way that pattern recognition is made with use of parts of a pattern being formed of a joint or loop. However, the pattern recognizing system disclosed in the Japanese Patent Publication No. 62-15911 has the disadvantage that if a continuously hand-written character to be recognized has many curves as in hiragana, elongated or contracted parts of the character make it difficult to designate in advance what parts of the character are, resulting in wrong recognition. The pattern recognizing system disclosed in the Japanese Patent Laid-Open No. 2-56689 has the disadvantage that a possible turbulence or distortion of the shape makes it difficult to judge what the shape is, resulting in wrong recognition.
SUMMARY OF THE INVENTION
The pattern data processing system of the present invention on the assumption of multi-user has the following three features.
The first feature of the present invention is a pattern data processing system, comprising: an input arrangement for receiving given patterns at least, a dictionary having data classified plurally for a plurality of users, a common dictionary, and a dictionary editing arrangement; wherein the dictionary has a group attribute of the users therein, if the pattern given through any of the input arrangements meets predetermined conditions the dictionary editing arrangement edits the data in the dictionary and enters the pattern and code corresponding to the pattern in the group dictionary, and the users having the same group attribute is accessible to the group dictionary.
In detail, the above-described first feature comprises an input arrangement having patterns entered therein, a dictionary having patterns, categories having the patterns belong thereto, and groups defined therein, an attribute designating arrangement for designating group attributes, such as user
1
, user
2
, group
1
, group
2
, and all members, a recognizing arrangement for receiving the entered pattern and group attribute before retrieving from among the patterns entered in the dictionary a pattern most similar to the entered one and having a corresponding group attribute to feed out the category, and a dictionary editing arrangement for extracting a pattern used in common by a group before editing the pattern, such as entering it in the common dictionary, with the group having the user made to belong thereto.
Operation of the first feature is described below. First, the entered pattern, such as a character, voice, or code string, entered through the input arrangement is fed to the recognizing arrangement together with the group attribute designated by the attribute designating arrangement. In turn, when receiving the entered pattern and the group attribute, the recognizing arrangement retrieves from among the patterns in the dictionary the pattern most similar to the entered one and having a corresponding group attribute before feeding out the category of the pattern belonging thereto. The category is a concept higher than the pattern and has at least one or more patterns belong thereto. If the entered pattern is a code string, the recognizing arrangement retrieving an entered pattern from among the patterns in the dictionary before feeding out a code indicating a process associated with the pattern. The dictionary editing arrangement checks contents of the dictionary before extracting a pattern that would be used in common by the group and enters the pattern with the group attribute of the pattern being the group. Therefore, even if the user belonging to the group does not enter the pattern by himself or herself, the first feature allows pattern recognition and pattern process with use of the pattern.
The second feature of the present invention is a pattern data processing system, comprising: a plurality of input arrangement for at least receiving given patterns, dictionaries for storing the entered patterns and one or a plurality of codes corresponding to t
Arai Toshifumi
Fukunaga Yasushi
Gunji Keiko
Kuzunuki Soshiro
Miura Masaki
Kenyon & Kenyon
Patel Jayanti K.
LandOfFree
Collaborative learning system and pattern recognition method does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Collaborative learning system and pattern recognition method, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Collaborative learning system and pattern recognition method will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2904165