Il-Seok Oh

Hybrid genetic algorithms for feature selection

This paper proposes a novel hybrid genetic algorithm for feature selection. Local search operations are devised and embedded in hybrid GAs to fine-tune the search. The operations are parameterized in terms of their fine-tuning power, and their effectiveness and timing requirements are analyzed and compared. The hybridization technique produces two desirable effects: a significant improvement in the final performance and the acquisition of subset-size control. The hybrid GAs showed better convergence properties compared to the classical GAs. A method of performing rigorous timing analysis was developed, in order to compare the timing requirement of the conventional and the proposed algorithms. Experiments performed with various standard data sets revealed that the proposed hybrid GA is superior to both a simple GA and sequential search algorithms.

A text watermarking algorithm based on word classification and inter-word space statistics

Text documents can be watermarked by patterning theinter-word spaces. This paper proposes a textwatermarking algorithm that exploits the novel conceptsof word classification and inter-word space statistics. Thewords are classified using some features. Severaladjacent words are grouped into a segment, and thesegments are also classified using the word classinformation. The same amount of information is insertedinto each of the segment classes. The information isencoded by modifying some statistics of inter-word spacesof the segments belonging to the same class. Severaladvantages over the conventional word-shift algorithmsare discussed.

A class-modular feedforward neural network for handwriting recognition

Abstract Since the conventional feedforward neural networks for character recognition have been designed to classify a large number of classes with one large network structure, inevitably it poses the very complex problem of determining the optimal decision boundaries for all the classes involved in a high-dimensional feature space. Limitations also exist in several aspects of the training and recognition processes. This paper introduces the class modularity concept to the feedforward neural network classifier to overcome such limitations. In the class-modular concept, the original K -classification problem is decomposed into K 2-classification subproblems. A modular architecture is adopted which consists of K subnetworks, each responsible for discriminating a class from the other K −1 classes. The primary purpose of this paper is to prove the effectiveness of class-modular neural networks in terms of their convergence and recognition power. Several cases have been studied, including the recognition of handwritten numerals (10 classes), English capital letters (26 classes), touching numeral pairs (100 classes), and Korean characters in postal addresses (352 classes). The test results confirmed the superiority of the class-modular neural network and the interesting aspects on further investigations of the class modularity paradigm.

A fast algorithm for tracking human faces based on chromatic histograms

For the applications like video conferencing, real-time tracking of human faces is an essential task. In this paper, we propose a fast algorithm for tracking human face regions in color motion images. This algorithm is based on chromatic histogram and histogram backprojection operation. Our experimental results show that the proposed algorithm can be used in real-time applications. Additionally it has the advantage of being insensitive to small variations of face regions.

Analysis of class separation and combination of class-dependent features for handwriting recognition

In this paper, we propose a new approach to combine multiple features in handwriting recognition based on two ideas: feature selection-based combination and class dependent features. A nonparametric method is used for feature evaluation, and the first part of this paper is devoted to the evaluation of features in terms of their class separation and recognition capabilities. In the second part, multiple feature vectors are combined to produce a new feature vector. Based on the fact that a feature has different discriminating powers for different classes, a new scheme of selecting and combining class-dependent features is proposed. In this scheme, a class is considered to have its own optimal feature vector for discriminating itself from the other classes. Using an architecture of modular neural networks as the classifier, a series of experiments were conducted on unconstrained handwritten numerals. The results indicate that the selected features are effective in separating pattern classes and the new feature vector derived from a combination of two types of such features further improves the recognition rate.

Distance features for neural network-based recognition of handwritten characters

Abstract. Features play an important role in OCR systems. In this paper, we propose two new features which are based on distance information. In the first feature (called DT, Distance Transformation), each white pixel has a distance value to the nearest black pixel. The second feature is called DDD (Directional Distance Distribution) which contains rich information encoding both the black/white and directional distance distributions. A new concept of map tiling is introduced and applied to the DDD feature to improve its discriminative power. For an objective evaluation and comparison of the proposed and conventional features, three distinct sets of characters (i.e., numerals, English capital letters, and Hangul initial sounds) have been tested using standard databases. Based on the results, three propositions can be derived to confirm the superiority of both the DDD feature and the map tilings.

Watermarking text document images using edge direction histograms

This paper proposes a novel watermarking algorithm for grayscale text document images. The algorithm inserts the watermark signals through edge direction histograms. The concept of sub-image consistency is developed. The concept means the sub-images have similar-shaped edge direction histograms and it is shown to be valid over a wide range of document images. Algorithms to insert and detect watermark signals are proposed. The experiments performed with various document images produced plausible results in terms of robustness.

Classifier ensemble selection using hybrid genetic algorithms

This paper proposes a hybrid genetic algorithm for classifier ensemble selection. In this paper, two local search operations used to improve offspring prior to replacement are proposed. The operations are parameterized in order to control the computation time. Experimental results and statistical tests demonstrate the effectiveness of the proposed hybrid genetic algorithm and related local search operations.

Local search-embedded genetic algorithms for feature selection

This paper proposes a novel hybrid genetic algorithm for the feature selection. Local search operations used to improve chromosomes are defined and embedded in hybrid GAs. The hybridization gives two desirable effects: improving the final performance significantly and acquiring control of subset size. For the implementation reproduction by readers, we provide detailed information of GA procedure and parameter setting. Experimental results reveal that the proposed hybrid GA is superior to a classical GA and sequential search algorithms.

Binary classification trees for multi-class classification problems

This paper proposes a binary classification tree aiming atsolving multi-class classification problems using binaryclassifiers. The tree design is achieved in a way that aclass group is partitioned into two distinct subgroups at anode. The node adopts the class-modular scheme toimprove the binary classification capability. Thepartitioning is formulated as an optimization problemand a genetic algorithm is proposed to solve theoptimization problem. The binary classification tree iscompared to the conventional methods in terms ofclassification accuracy and timing efficiency.Experiments were performed with numeral recognitionand touching-numeral pair recognition.