VLSI neural fuzzy classifier for handwriting recognition

• US 20040008883A1
• Filed: 07/12/2002
• Published: 01/15/2004
• Est. Priority Date: 07/12/2002
• Status: Active Grant

• Current Assignee: Winbond Electronics Corporation
• Original Assignee: Winbond Electronics Corporation
• Inventors: Li, Guoxing, Shi, Bingxue
• Granted Patent: US 7,146,037 B2
• Time in Patent Office: Days
• US Class Current: 382/158
• CPC Class Codes: G06F 18/24 Classification techniques
• CPC Class Codes: G06V 10/955 using specific electronic processors
• CPC Class Codes: G06V 30/10 Character recognition

Abstract

A handwriting recognition device using fuzzy logic and cellular neural network for unconstrained handwritten numeral classification is provided. The current mode VLSI classifier has a I/O circuit for inputting and outputting a plurality of membership functions. An extraction unit comprising a CCD extractor with a CNN structure and a compression unit receives a to-be-recognized character having a plurality of input features for generating a plurality of features values after compression. A membership function generator stores the plurality of membership functions and receives the plurality of features values to generate a plurality of current-type membership degrees. A plurality of switched-current integrators receives the plurality of current-type membership degrees for generating a plurality of synthesis membership degrees. A k-WTA circuit is provided for comparing the plurality of synthesis membership degrees and output the plurality of synthesis membership degrees as well as the corresponding characters in an order of magnitude.

Claims

1. A handwritten numeral classifier using fuzzy logic and cellular neural network, comprising:

an extraction unit using cellular neural network for receiving a scanned image having a plurality of input features, and compressing the received data of the scanned image to generate a plurality of feature values;

a membership function generator using fuzzy logic for storing a plurality of membership functions and receiving the feature values to generate a plurality of synthesis membership function degrees for the plurality of input features;

a k-WTA circuit for receiving the plurality of synthesis membership function degrees from the membership function generator and outputting the synthesis membership degrees in order of magnitude;

an I/O circuit for inputting programming codes to the membership function generator through off-chip memory units and receiving the synthesis membership degrees from the k-WTA circuit to output a final recognizing result of the scanned image; and

a clock generator and logic controller for generating clock cycle and control logic signals for controlling timing of and logic for operations of the extraction unit, membership function generator, and the k-WTA circuit.

2. The classifier in claim 1, wherein the I/O circuit and the k-WTA circuit both have 11 corresponding ports.

3. The classifier in claim 1, wherein the membership function generator further comprising:

a membership function generator array respectively corresponding to each one of the plurality of the input features for storing the plurality of membership functions and generating a plurality of current-type membership function degrees for the plurality of the input features, and

a plurality of switched-current integrator corresponding to the membership function generator array for receiving the plurality of membership function degrees and generating a plurality of synthesis membership degrees.

4. The classifier in claim 3, wherein the membership function generator array is 10×

10.

5. The classifier in claim 3, wherein the number of SI integrator includes 10.

6. The classifier in claim 5, wherein each of the plurality of SI integrators includes a plurality of storage units constructed by a RegulatedGate Cascade structure.

7. The extraction unit in claim 1, wherein the extraction unit further comprising:

a connected components detector extractor with a cellular neural network structure for speedily extracting the input features of a scanned numeral image; and

a compression unit for compressing bits of the input features into small and meaningful feature values and sending the compressed data to the membership function generator.

8. The classifier in claim 7, wherein the CCD extractor is 24 bits.

9. The classifier in claim 1, wherein the k-WTA circuit and the I/O circuit have 11 corresponding ports.