An image-forming device has a plurality of cylindrical image-carrying members, and an optical scanning unit. The plurality of cylindrical image-carrying members is provided in a main casing for rotating about respective rotational axes. The rotational axes are parallel to one another and juxtaposed in a single direction. The optical scanning unit is provided in the main casing and has a plurality of laser generators, a single rotatable polygon mirror, a plurality of mirrors, and a plurality of lenses. The plurality of laser generators is provided one-to-one relationship to the plurality of image-carrying members to emit respective laser beams. The single rotatable polygon mirror deflects the respective laser beams to scan the respective image-carrying members therewith. The plurality of mirrors is provided one-to-one relationship to the plurality of image-carrying members to guide the respective laser beams along respective optical paths to the respective image-carrying members. The plurality of lenses is provided one-to-one relationship to the plurality of image-carrying members and provided in the respective optical paths and between the respective mirrors and the respective image-carrying members. Distances from the respective mirrors to the respective image-carrying members being equal to one another.

Image-forming device

PURPOSE:To correctly recognize a character even in the character having a very similar shape by deciding the character needing a correction and the character type after the correction from the arrangement of the character type of respective first candidate characters, and making the recognizing candidate character having the same character type as the character after the correction from the recognizing candidate character group of the character needing the correction into a first candidate character. CONSTITUTION:A distance counting part 5 obtains the distance of the characteristic quantity of a recognizing object character pattern obtained by a characteristic extracting part 4 and the characteristic quantity of respective characters registered in a first dictionary 6 beforehand. A classifying part 7 makes the smallest distance out of the distance obtained by the distance counting part 5 into the first candidate character, makes the N character into the recognizing candidate character group from the character with a small distance including the character and adds the character type to display the attribute of respective characters to respective recognizing candidate characters. In the character where the character type is changed by a character type changing part 8, a correcting part 9 makes the highest order candidate character out of the candidate character having the same character type as the character type after the changing of the character to change the character type in a second to the N-th candidate characters into the first candidate character.

Character recognizing device

A vacuum drum apparatus on which a mask is mounted having a resilient surface and an array of holes suitable for a substrate of a given size. The drum is a circular cylinder perforated with an array of holes. A hollow chamber within the drum is connected to a controlled pressure source for providing a vacuum. The mask may be magnetically attracted about the outer surface of the drum. Thus, the mask may be made of a flexible ferromagnetic sheet. The mask is preferably elastomeric and may include an inextensible layer.

Variable mask and universal vacuum drum

PURPOSE:To improve the recognition rate by removing an abnormal reflected light area from the object to perform comparison and collation at the time of the existence of this area in a scan area and removing a drop-out part from the object to perform coincidence discrimination at the time of the occurrence of partial drop-out in a read input character pattern due to an interposed transparent medium. CONSTITUTION:If the abnormal reflected light area exists, the signal level in an area X of an electric signal is higher than the normal paper surface level and is outputted to a binarizing circuit 20 and an abnormal reflected light area detecting circuit 50, and a binarized picture signal is outputted to a character extracting circuit 30. The character extracting circuit 30 normalizes a character to a fixed size by the inputted picture signal and inputs the result to one input of a collating and discriminating circuit 40. The abnormal reflected light area detecting circuit 50 binarizes the signal based on the paper surface level to detect the abnormal reflected light area and outputs it to an abnormal reflected area position detecting circuit 60 to determine whether the abnormal reflected light area exists or not. The collating and discriminating circuit 40 discriminates the character by pattern matching between standard character patterns and the normalized signal, and the area position is removed from the object of pattern matching. Thus, the recognition rate is improved.

Optical character reader

The signal to be classified is sampled and the samples are multiplied by weighting functions prior to performing discrete Fourier transforms, power calculations, and normalization thereon so that the preprocessor is essentially a plurality of channel bandpass filter stages, the k-th one of which has a frequency response approximating an ideal filter defined by ##EQU1## where k=O, α, 2α, 3α, . . . , rα α=constant, 1≦α≦N/2 r+1=greatest integer in N/α This preprocessor transforms blocks of one dimensional data into spectra in such a way that data blocks coming from similar sources will have spectra that are close to one another in spectrum space. The discrete Fourier transform is modified to remove or reduce the phase dependency problem and to enhance the clustering of similar spectra by the weighting function above.

Preprocessor for spectral pattern classification systems

Character input device

PURPOSE:To attain separation/extraction of characters which are written with no writing conditions by tracking the outer circumference of a character within a picture of an input character string with fixed conditions and segmenting an area of a character including the character to eliminate noises other than the character. CONSTITUTION:A character string area within a picture is fetched to a buffer memory 2 of a rectangular area after a digital picture such as a logical circuit diagram, etc. is stored to a picture memory 1. Then the character string area is scanned vertically to detect a tracking start point for outer circumference of character. Thus the outer circumference of the character is tracked by a round based on said start point and in a fixed checking order. A character frame area is segmented according to the minimum and maximum points which are extracted from said tracking process and stored to a one-character buffer memory 5 and then to a one-character store part 7 after elimination of noises through a noise eliminating part 6. The stored frame area is used for recognition of characters and also supplied to a one-character erasing part 8, and an EOR operation is performed between the separated character and a corresponding character of the memory 2. Then a character string including no processed character is stored to the memory 2.

Character separating system

PURPOSE:To extract correctly an end point out of a linear pattern in the form of a pseudo land, by using a pseudo end point mark extracting part containing a specific contour code producing part and direction code producing part. CONSTITUTION:An observing frame 8 of (nXm)-mesh set centering on the center of an intersection 7 of grating lines is cut out, and a pseudo end point mark extracting part is provided to extract the end point of a linear pattern 2. This mark extracting part includes a part which produces a contour code that decides the intersecting position of the pattern 2 among 4 sides and 4 corners of the frame 8. At the same time, the inside of the frame 8 is scanned by a mesh window 12 of (pXq)-mesh to extract the 4 directional components of the lateral, longitudinal, right lowering and left lowering directions respectively. Then a direction code is produced based on the state of existence of these directional components.

Extracting and processing system for graphic information

PURPOSE:To process a loop pattern having no cross points, by erasing segment patterns other than loop patterns and extracting artificial end points of loop patterns to trace them. CONSTITUTION:A pattern extracting part 41 converts linear patterns to labels in lattice units. An end point code extracting circuit 42 extracts end point codes like end codes. An end point extracting circuit 43 extracts end points of segments having no end point codes and sends them to an integrating circuit 44. In respect to picture data set in a table memory 45, patterns other than loop patterns having no cross points are erased by a cross point extracting circuit 46 and an end point tracing processing circuit 47. When a loop pattern having no cross points is left, the start point is obtained by an artificial end point extracting circuit 48, and the pattern is traced in a tracing processing circuit 49 49 by the known method.

Processing system for extraction of pattern information

PURPOSE:To cope with character cut out of a handwritten drawing, a free format document, OCR paper without character frame etc. by enabling to separate character one by one even when characters are in contact with each other in the data. CONSTITUTION:In a character data processing device, the result of optical scanning of a drawing in which handwritten characters to be segmented are written is stored in a picture memory 1, and scanned data are loaded from inside of the picture memory 1. Character area of the loaded data is determined by a character area determining section 2. A character frame of the determined character area is extracted, and segmenting of individual character is made by a character frame extracting section 3. Segmenting of individual character is identified supposing that, for instance, the width W of a character is within 2/3h where (h) is the height of the character. That is, the characters adjoining but written off from each other are segmented on the basis of such width W.

Matsuura Toshio

Papers

Character input device

Character separating system

Extracting and processing system for graphic information

Processing system for extraction of pattern information

Character data processing device