The multiplierless fast Fourier transform (FFT) with dyadic-valued (rational) coefficients is important for many signal processing tools. The proposed lifting based FFT (L-FFT) based on fast Hartley transform (FHT) has a simpler structure than existing ones because fewer lifting steps need to be approximated. In addition, it has a structure of real-valued calculation followed by complex-valued parts, thereby it requires fewer memories for the internal implementation than the conventional FFTs.

Discrete cosine transform (DCT) is known as efficient frequency transform, and when it is implemented on software/hardware, multiplier is undesirable for faster implementation. This paper presents a realization of multiplierless fast DCT for lossy image/video coding on arbitrary devices. First, the proposed DCT is constructed by using fast Hartley transform (FHT). Next, the redundancy of the structure is eliminated by using several characteristics of rotation matrix. Then, multiplierless DCT is obtained by approximating rotation matrices to multiplierless lifting structures with adders and bit-shifters. Finally, the proposed DCT is validated by comparing with the conventional DCTs in image coding.

There exist distributed processing environments composed of many heterogeneous computers. It is required to schedule distributed parallel processes in an appropriate manner. For the scheduling, prediction of execution load of a process is effective to exploit resources of environments. We propose long-term load prediction methods with references of properties of processes and of runtime predictions. Since an appropriate prediction method is different according to the situation, we propose a prediction module selection to select an appropriate prediction method according to a state of changing CPU load using a neural network. We also discuss about the implementation of a long-term CPU load prediction system, which provides information including prediction of load for schedulers, system administrators and users.

Conventional segmentation technique cannot extract difficult characters such as an isolated character and a touching character. In this paper, we propose a novel character recognition method which executes segmentation and recognition simultaneously. This method enables us to extract and recognize such difficult characters. The effectiveness of the proposed method is confirmed by experiments.

Decorated characters are widely used in various documents. Practical optical character reader is required to deal with not only common fonts but also complex designed fonts. However, since the appearances of decorated characters are complicated, most general character recognition systems cannot give good performances on decorated characters. In this paper, an algorithm that can extract character's essential structure from a decorated character is proposed. This algorithm is applied in preprocessing of character recognition. The proposed algorithm consists of three procedures: global structure extraction, interpolation of structure and smoothing. By using multiscale images, topographical features, such as ridges and ravines are detected for structure extraction. Ridges are used for extracting global structure and ravines are used for interpolation. Experimental results show character structures can be clearly extracted from very complex decorated characters.

This paper proposes a parallel architecture for quadtree-based fractal image coding. This architecture is capable of performing the fractal image coding based on quadtree partitioning without the external memory for the fixed domain pool. Since a large domain block consists of small domain blocks, the calculations of distortion for all kinds of domain blocks are performed by the summation of the distortions for the maximum-depth domain pool which is extracted from the smallest range blocks of the neighbor processors. Fast comparison module is proposed for this architecture. This module can compute the distortions between range blocks and their eight isometric transformations by one full rotation around the center.

Some methods have been developed for character extraction from document images and scene images with complex backgrounds. However, those methods are designed to extract rather large characters, and often fails to extract small characters. This paper proposes a new method by which character patterns can be extracted from document images with complex background. The method is based on the local multilevel thresholding and pixel labeling, and the region growing. This framework is very useful for extracting character patterns from badly illuminated document images. The performance of extracting small character patterns has also been improved by suppressing the influence of mixed-color pixels around character's edges.

In statistical pattern recognition, parameters of distributions are usually estimated from training samples. It is well known that shortage of training samples causes estimation errors which reduce recognition accuracy. By studying estimation errors of eigenvalues, various methods of avoiding recognition accuracy reduction have been proposed. However, estimation errors of eigenvectors have not been considered enough. In this paper, we investigate estimation errors of eigenvectors to show these errors are another factor of recognition performance reduction. We propose a new method for modifying eigenvalues in order to reduce bad influence caused by estimation errors of eigenvectors. Effectiveness of the method is shown by experimental results.

Distorted character recognition is a difficult but in-evitable problem in hand-printed character recognition. In this paper, we propose a character recognition method us-ing elastic models for recognizing cursive characters with intricate structure. The models are fitted to unknown in-put patterns by applying the EM algorithm to minimize a measure of fittness. To avoid falling into local minima, mul-tiresolutional approach is introduced. Moreover, nonlinear transformation is adopted to realize more flexible matching. Experiments performed on Japanese characters show effec-tiveness of the proposed method. 1.

Decorated characters are widely used in various documents. Practical optical character reader is required to deal with not only common fonts but also complex designed fonts. However, since appearances of decorated characters are complicated, most general character recognition systems cannot give good performances on decorated characters. In this paper, an algorithm that can extract character's essential structure from a decorated character is proposed. This algorithm is applied in preprocessing of character recognition. The proposed algorithm consists of three parts: global structure extraction, interpolation of structure, and smoothing. By using multi-scale images, topographical features such as ridges and ravines are detected for structure extraction. Ridges are used for extracting global structure, and ravines are used for interpolation. Experimental results show clear character structures are extracted from very complex decorated characters.

For many pattern recognition methods, high recognition accuracy is obtained at very high expense of computational cost. In this paper, a new algorithm that reduces the computational cost for calculating discriminant function is proposed. This algorithm consists of two stages which are feature vector. Division and dimensional reduction. The processing of feature division is based on characteristic of covariance matrix. The dimensional reduction in the second stage is done by an approximation of the Mahalanobis distance. Compared with the well-known dimensional reduction method of K-L expansion, experimental results show the proposed algorithm not only reduces the computational cost but also improves the recognition accuracy.

Accurate recognition of blurred images is a practical but previously mostly overlooked problem. In the paper, we quantify the level of noise in blurred images and propose a modification of discriminant functions that adapts to the level of noise. Experimental results indicate that the proposed method actually enhances the existing statistical methods and has impressive ability to recognize blurred image patterns.

Statistic language models obtained from a large number of training samples play an important role in speech recognition. In order to obtain higher recognition performance, we should introduce long distance correlations between words. However, traditional statistic language models such as word n-grams and ergodic HMMs are insufficient for expressing long distance correlations between words. We propose an acquisition method for a language model based on HMnet taking into consideration long distance correlations and word location.

This paper presents a precise system for handwritten Chinese and Japanese character recognition. Before extracting directional element feature (DEF) from each character image, transformation based on partial inclination detection (TPID) is used to reduce undesired effects of degraded images. In the recognition process, city block distance with deviation (CBDD) and asymmetric Mahalanobis distance (AMD) are proposed for rough classification and fine classification. With this recognition system, the experimental result of the database ETL9B reaches to 99.42%.

Many Japanese documents contain both horizontally and vertically written text lines, and furthermore, those text lines often have local skew. In this paper, we propose a framework for detecting the local orientation of text lines and selecting text line candidates of correct orientation. The orientation detection is independent of text block extraction and has advantage of being applicable to complicated documents. The framework is intended to reduce the text line candidates of incorrect orientation and to improve the performance of document image analysis systems.

An algorithm to construct a multi-template dictionary for character recognition is presented. It is a repetitive process of partitioning and estimating the distribution by observing both within-class distribution and between-class information. The effectiveness of the algorithm is shown by experimental results.

The tree-structured speaker clustering was proposed as a highspeed speaker adaptation method. It can select the model which is most similar to a target speaker. However, this method does not consider speaker difference dependent on phoneme class. In this paper, we propose a speaker adaptation method based on speaker clustering by taking speaker difference dependent on phoneme class into account. The experimental results showed that the new method gave a better performance than the original method. Furthermore, we propose the improved method which use a tree-structure of a similar phoneme as the substitute for the phoneme which does not appear in the adaptation data. From the experimental results, the improved method gave a better performance than the method previously proposed.

N-gram model is widely applied to various pattern recognition system because it well represents local features of natural languages. In this paper, we describe a contextual postprocessing method using a trigram model of character for Japanese document recognition, and its advantage is revealed by practical experiments. The model is automatically obtained by statistical processing of training documents. The ability to reduce ambiguity is evaluated by the perplexity. In the processing, two smoothing methods are examined, and the predictive power of the deleted interpolation method is shown to be superior. For leading articles, the perplexity reduced to about 22 when using deleted interpolation. The output from OCR is processed very fast using a Viterbi algorithm. Experimental results of recognition for three kinds of documents show that the error correction rates are ranged from 75 to over 90 percent.

An intelligent character recognition system with high accuracy and high speed is realized by integrating image-type and logical-type information processing modules. For image-type information processing, a technique called associative pattern matching is proposed, and its usefulness is verified by experiments. The results show that it is useful for rough classification of input patterns. For logical-type processing, structural analysis of characters is realized by a relaxation matching technique, which is described in detail, and some experimental results are shown. The integration of these two techniques is discussed.>

A general fixed structure automaton acting in stationary random multireaction environment is described. Some necessary conditions are shown for automata to be purposeful or adaptive in an unknown environment (to be more profitable than a no-purpose, or random, automaton). By the necessary conditions it is proved that there exists no perfectly adaptive automaton. A necessary condition for automata with minimum state space to be purposeful is obtained. This condition characterizes their structures explicitly, so that a little restriction gives a sufficient condition. Some general subclasses of automata are proposed, and necessary and sufficient conditions for such automata to be purposeful are investigated. Furthermore as properties of a purposeful automaton, the time required to adapt and the stability of adapting to the random environment are discussed.