Character recognition apparatus

• US 4,607,385 A
• Filed: 05/22/1984
• Issued: 08/19/1986
• Est. Priority Date: 05/25/1983
• Status: Expired due to Term

• Current Assignee: Kabushiki Kaisha Toshiba (Toshiba Corporation)
• Original Assignee: Kabushiki Kaisha Toshiba (Toshiba Corporation)
• Inventors: Maeda, Ken-ichi
• Primary Examiner(s): NOT, DEFINED
• Application Number:
  US06/613,069
• Time in Patent Office: 819 Days
• Field of Search: 382/9, 382/34, 382/66
• US Class Current: 382/179
• CPC Class Codes: G06V 10/10 Image acquisition document image scanning and transmission H04N1/00; control of digital cameras H04N23/60
• CPC Class Codes: G06V 30/10 Character recognition
• CPC Class Codes: G06V 30/146 Aligning or centring of the image pick-up or image-field
• CPC Class Codes: G06V 30/153 using recognition of characters or words
• CPC Class Codes: G06V 30/192 using simultaneous comparisons or correlations of the image signals with a plurality of references

Abstract

A character recognition apparatus has an image data generating section for storing image data corresponding to a character array on a sheet of paper and for generating a series of scan line data constituting one item of image data; a segmentation control section for generating starting and end scan line data which define a segment of each character data in accordance with the scan line data sequentially generated from the image data generating section; and a recognizing section for detecting any similarities between the reference character data and the character data generated from the data generating section in accordance with the starting and end scan line data and for generating effective data character data which has the largest similarity which is larger than a predetermined value. When none of the similarities exceed the predetermined value, the recognizing section supplies a control signal to the segmentation control section to update one item of the starting and end scan line data, thereby allowing the image data generating section to generate character data for a different segment.

Claims

1. A character recognition apparatus comprising:

scan line data generating means for detecting at least one character on a medium, storing image data corresponding to a detected character, and generating a series of scan line data constituting the image data;

segmentation control means for generating starting and end scan line position data which respectively represent starting and end scan line positions for defining a segment of each character data in accordance with the series of scan line data;

character selecting means for generating from said scan line data generating means the character data specified in accordance with the starting and end scan line position data generated from said segmentation control means; and

recognizing means for recognizing characters by comparing the character data of the segment defined by the starting and end scan line position data generated from said segmentation control means with a plurality of prestored reference character data, and for supplying a control signal to said segmentation control means to update at least one of the starting and end scan line position data when the character data from said scan line data generating means does not coincide with any of the reference character data.

2. An apparatus according to claim 1, wherein said scan line data generating means comprises a charge-coupled device scanner for optically detecting characters, an analog-to-digital converter for converting to a digital signal an output signal generated from said charge-coupled device scanner, and a memory for storing the digital signal generated from said analog-to-digital converter.

3. An apparatus according to claim 2, wherein said segmentation control means comprises a boundary signal generating circuit for receiving each of the scan line data generated from said scan line data genearting means and generating a first signal which represents a change from a noncharacter area to a character area and a second signal which represents a change from the character area to the noncharacter area in accordance with the received scan line data, a counting circuit for counting the scan line data supplied from said scan line data generating means to said boundary signal generating circuit, first and second data storing means for respectively storing as the starting and end scan line position data contents of said counting circuit in response to the first and second signals generated from said boundary signal generating circuit, and readout control means, reponsive to the control signal generated from said recognizing means, for updating at least one of read address signals supplied to said first and second data storing means to update at least one of the starting and end scan line position data from said first and second data storing means, and supplying updated starting and end scan line position data to said character selecting means.

4. An apparatus according to claim 3, wherein said recognizing means comprises a data generator for generating similarity data corresponding to similarities between a plurality of prestored reference character data and character data generated from said scan line data generating means in accordance with the starting and end scan line position data, a memory for storing similarity data whose similarities exceed a predetermined value, and an interpretation circuit for generating as effective character data the character data associated with largest similarity data stored in said memory and for generating the control signal when said interpretation circuit detects that none of similarity data exceeding the predetermined value is stored in said memory.

5. An apparatus according to claim 4, wherein said data generator comprises a normalization circuit for normalizing the character data generated from said scan line data generating means, a blurring circuit for blurring the character data generated from said normalization circuit, a reference memory for storing the plurality of prestored reference character data, and a similarity calculator for calculating a similarity between the character data generated from said blurring circuit and the reference character data read out from said reference memory.

6. An apparatus according to claim 3, wherein said recognizing means comprises a data generator for generating similarity data corresponding to similarities between a plurality of prestored reference character data and character data generated from said scan line data generating means in accordance with the starting and end scan line position data, a memory for storing similarity data whose similarities exceed a predetermined value, a word-similarity calculator for calculating a word-similarity corresponding to similarities between word data constituted by the character data of a series of segments stored in said memory and a plurality of prestored reference word data, and an interpretation circuit for generating as effective data the word data associated with largest word-similarity data generated from said word-similarity calculator.

7. An apparatus according to claim 6, wherein said data generator comprises a normalization circuit for normalizing the character data generated from said scan line data generating means, a blurring circuit for blurring the character data generated from said normalization circuit, a reference memory for storing the plurality of prestored reference character data, and a similarity calculator for calculating a similarity between the character data generated from said blurring circuit and the reference character data read out from said reference memory.

8. An apparatus according to claim 2, wherein said recognizing means comprises a data generator for generating similarity data corresponding to similarities between a plurality of prestored reference character data and character data generated from said scan line data generating means in accordance with the starting and end scan line position data, a memory for storing similarity data whose similarities exceed a predetermined value, and an interpretation circuit for generating as effective character data the character data associated with largest similarity data stored in said memory and for generating the control signal when said interpretation circuit detects that none of similarity data exceeding the predetermined value is stored in said memory.

9. An apparatus according to claim 2, wherein said recognizing means comprises a data generator for generating similarity data corresponding to similarities between a plurality of prestored reference character data and character data generated from said scan line data generating means in accordance with the starting and end scan line position data, a memory for storing similarity data whose similarities exceed a predetermined value, a word-similarity calculator for calculating a word-similarity corresponding to similarities between word data constituted by the character data of a series of segments stored in said memory and a plurality of prestored reference word data, and an interpretation circuit for generating as effective data the word data associated with largest word-similarity data generated from said word-similarity calculator.

10. An apparatus according to claim 1, wherein said scan line data generating means comprises a flying spot scanner for optically detecting characters and storing detected character data, and an analog-to-digital converter for converting to a digital signal an output signal generated from said flying spot scanner.

11. An apparatus according to claim 10, wherein said segmentation control means comprises a boundary signal generating circuit for receiving each of the scan line data generated from said scan line data genearting means and generating a first signal which represents a change from a noncharacter area to a character area and a second signal which represents a change from the character area to the noncharacter area in accordance with the received scan line data, a counting circuit for counting the scan line data supplied from said scan line data generating means to said boundary signal generating circuit, first and second data storing means for respectively storing as the starting and end scan line position data contents of said counting circuit in response to the first and second signals generated from said boundary signal generating circuit, and readout control means, reponsive to the control signal generated from said recognizing means, for updating at least one of read address signals supplied to said first and second data storing means to update at least one of the starting and end scan line position data from said first and second data storing means, and supplying updated starting and end scan line position data to said character selecting means.

12. An apparatus according to claim 11, wherein said recognizing means comprises a data generator for generating similarity data corresponding to similarities between a plurality of prestored reference character data and character data generated from said scan line data generating means in accordance with the starting and end scan line position data, a memory for storing similarity data whose similarities exceed a predetermined value, and an interpretation circuit for generating as effective character data the character data associated with largest similarity data stored in said memory and for generating the control signal when said interpretation circuit detects that none of similarity data exceeding the predetermined value is stored in said memory.

13. An apparatus according to claim 11, wherein said recognizing means comprises a data generator for generating similarity data corresponding to similarities between a plurality of prestored reference character data and character data generated from said scan line data generating means in accordance with the starting and end scan line position data, a memory for storing similarity data whose similarities exceed a predetermined value, a word-similarity calculator for calculating a word-similarity corresponding to similarities between word data constituted by the character data of a series of segments stored in said memory and a plurality of prestored reference word data, and an interpretation circuit for generating as effective data the word data associated with largest word-similarity data generated from said word-similarity calculator.

14. An apparatus according to claim 10, wherein said recognizing means comprises a data generator for generating similarity data corresponding to similarities between a plurality of prestored reference character data and character data generated from said scan line data generating means in accordance with the starting and end scan line position data, a memory for storing similarity data whose similarities exceed a predetermined value, and an interpretation circuit for generating as effective character data the character data associated with largest similarity data stored in said memory and for generating the control signal when said interpretation circuit detects that none of similarity data exceeding the predetermined value is stored in said memory.

15. An apparatus according to claim 10, wherein said recognizing means comprises a data generator for generating similarity data corresponding to similarities between a plurality of prestored reference character data and character data generated from said scan line data generating means in accordance with the starting and end scan line position data, a memory for storing similarity data whose similarities exceed a predetermined value, a word-similarity calculator for calculating a word-similarity corresponding to similarities between word data constituted by the character data of a series of segments stored in said memory and a plurality of prestored reference word data, and an interpretation circuit for generating as effective data the word data associated with largest word-similarity data generated from said word-similarity calculator.

16. An apparatus according to claim 1, wherein said segmentation control means comprises a boundary signal generating circuit for receiving each of the scan line data generated from said scan line data genearting means and generating a first signal which represents a change from a noncharacter area to a character area and a second signal which represents a change from the character area to the noncharacter area in accordance with the received scan line data, a counting circuit for counting the scan line data supplied from said scan line data generating means to said boundary signal generating circuit, first and second data storing means for respectively storing as the starting and end scan line position data contents of said counting circuit in response to the first and second signals generated from said boundary signal generating circuit, and readout control means, reponsive to the control signal generated from said recognizing means, for updating at least one of read address signals supplied to said first and second data storing means to update at least one of the starting and end scan line position data from said first and second data storing means, and supplying updated starting and end scan line position data to said character selecting means.

17. An apparatus according to claim 16, wherein said recognizing means comprises a data generator for generating similarity data corresponding to similarities between a plurality of prestored reference character data and character data generated from said scan line data generating means in accordance with the starting and end scan line position data, a memory for storing similarity data whose similarities exceed a predetermined value, and an interpretation circuit for generating as effective character data the character data associated with largest similarity data stored in said memory and for generating the control signal when said interpretation circuit detects that none of similarity data exceeding the predetermined value is stored in said memory.

18. An apparatus according to claim 1, wherein said recognizing means comprises a data generator for generating similarity data corresponding to similarities between a plurality of prestored reference character data and character data generated from said scan line data generating means in accordance with the starting and end scan line position data, a memory for storing similarity data whose similarities exceed a predetermined value, and an interpretation circuit for generating as effective character data the character data associated with largest similarity data stored in said memory and for generating the control signal when said interpretation circuit detects that none of similarity data exceeding the predetermined value is stored in said memory.

19. An apparatus according to claim 1, wherein said recognizing means comprises a data generator for generating similarity data corresponding to similarities between a plurality of prestored reference character data and character data generated from said scan line data generating means in accordance with the starting and end scan line position data, a memory for storing similarity data whose similarities exceed a predetermined value, a word-similarity calculator for calculating a word-similarity corresponding to similarities between word data constituted by the character data of a series of segments stored in said memory and a plurality of prestored reference word data, and an interpretation circuit for generating as effective data the word data associated with largest word-similarity data generated from said word-similarity calculator.