Research on some key techniques of offline handwritten Chinese character recognition
For the problem of large character set recognition, the algorithm of template matching is generally used, mainly because the algorithm is simple and the recognition speed is fast. However, the direct template matching algorithm is often unable to meet the needs of the recognition precision in practical applications. Based on the template matching algorithm and the principle of statistical analysis and statistical signal processing, this paper studies the key technologies of offline handwritten Chinese character recognition (Jinjunling), and makes a study on the algorithm of offline handwritten Chinese character recognition and its related problems, and tries to improve the recognition accuracy on the basis of not reducing the recognition speed.
Content Introduction Books
Computer Books
Research on some key technologies of offline handwritten Chinese character recognition from offline handwritten Chinese character recognition to the problem of large class number pattern recognition, in order to improve the precision of recognition based on the basic non-reduction of recognition speed, the establishment of handwritten Chinese character library, the construction of offline Chinese character recognition system, The character description and sample selection of Chinese characters in Chinese character recognition algorithm based on statistical analysis are described in detail.
Catalogue the 1th Chapter introduction
1. 1 issues raised
1. 2 Research Status Analysis
1. 3 Research content and main work
1. 4 structure and arrangement of the book
Reference documents
Chapter 2nd realization of offline handwritten Chinese character recognition system
2. 1 offline handwritten Chinese character recognition system
2. 1. 1 Recognition principle
2. 1. 2 class Two offline handwritten Chinese character recognition system based on template matching
2. 2 feature extraction of handwritten Chinese character recognition system
2. 2. 1 Definition of directional line
2. 2. Extraction of 2 directional line attributes
2. 3 recognition algorithm based on template matching
2. Coarse classification algorithm of 4 Chinese character recognition system
2. 4. 1 Rough classification Feature extraction
2. 4. 2 Recognition algorithm for coarse classification
2. A fine classification algorithm of 5 Chinese character recognition system
2. 5. 1 Fine classification Feature extraction
2. 5. 2 Recognition algorithm for fine classification
2. 6 Summary of this chapter
Reference documents
3rd Chapter HCL2004 offline handwritten Chinese character library and related research
3. 1 Research background and current situation
3. 2 HCL2000 handwritten Chinese character database
3. 2. 1 Database System Model
3. 2. 2 organization of Chinese character sample information
3. 2. 3 Management of Writers ' information
3. 2. Mutual search method of 42 kinds of information
3. 2. Data distribution of 5 HCL2000
3. 3 reasons for updating the HCL2000 database
3. HCL2004 system model and implementation of 4 handwritten Chinese character database
3. 4. 1 Forms of Chinese character sample information
3. 4. 2 Partitioning of sample sets
3. 4. Implementation of 3 HCL2004 handwritten Chinese character database
3. 5 analysis based on the HCL2004 database
3. 5. 1 Experimental system
3. 5. 2 Training sample count and recognition rate
3. 5. 3 Selection and identification of sample quality
3. 5. 4 performance analysis based on word recognition
3. 5. 5 analysis on the performance of distance measure classifier
3. 6 Summary of this chapter
Reference documents
The 4th Chapter research on handwritten Chinese character recognition algorithm based on statistical analysis
4. 1 Introduction
4. 2 several commonly used averages
4. 2. 1 mean value
4. 2. 2 Middle Digits
4. 3 Description of Dispersion degree of sample data
4. 3. 1 Standard deviation
4. 3. 2 very poor
4. Analysis of sample characteristics of 4 HCL2004 database
4. 5 standard handwritten Chinese character template based on average
4. 5. 1 Standard template based on mean value
4. 5. 2 Standard template based on the number of bits
4. 6 distance measure of inductive data dispersion degree parameter
4. 6. 1 introduction of a very poor distance measure
4. 6. 2 Introducing the distance measure of standard deviation
4. 7 experiments
4. 7. 1 different standard template classification performance analysis
4. 7. 2 analysis of classification performance of distance measure with different dispersion degree parameters
4. 8 Summary of this chapter
Reference documents
The 5th chapter is based on the high-order statistic distance measure
5. 1 Introduction
5. 2 introduction of higher-order statistics in distance measurement
5. 3 Distance measurement based on second-order standard deviation
5. 3. Definition of 12-order standard deviation
5. 3. 2 feasibility analysis of characterizing feature distribution with second-order standard deviation
5. 3. 3 Distance measurement based on second-order standard deviation
5. 3. 4 experiments
5. 4 distance measurement based on high-order statistics
5. 4. 1 3 kinds of high-order statistics
5. 4. 2 distance measurement based on high-order statistics
5. 4. 3 experiments
5. 5 Summary of this chapter
Reference documents
The 6th chapter of multi-level Chinese character recognition system based on sample clustering
6. 1 Introduction
6. 2 K-mean clustering algorithm based on DB criterion
6. 2. 1 k mean value algorithm
6. 2. 2 DB Validity criteria
6. 2. 3 K-Mean algorithm based on DB criterion
6. 3 multi-template matching algorithm
6. 3. Principle of 1 multi-template matching algorithm
6. 3. Design scheme of 2 multi-template matching algorithm
6. 4 experiments
6. 4. 1 Experimental system
6. 4. 2 system implementation
6. 4. 3 Experimental results and analysis
6. 5 Summary of this chapter
Reference documents
The 7th Chapter selection algorithm based on generalized confidence degree
7. 1 Introduction
7. Confidence analysis of 2 character recognition
7. 2. 1 reliability and generalized confidence of classifiers
7. 2. 2 Confidence estimates for classifiers
7. 3 definition of boundary samples based on generalized confidence level
7. 4 Sample selection algorithm based on generalized confidence degree
7. 5 Experimental results and analysis
7. 6 Summary of this chapter
Reference documents
8th Chapter Concluding remarks
< research on some key technologies of offline handwritten Chinese character recognition >
