Yoshihiro Shima

Document retrieval tolerating character recognition errors : Evaluation and application

Abstract This paper presents two methods of combining character recognition with techniques for retrieving Japanese documents and also shows how these methods can be applied to textual image retrieval. Both retrieval methods are tolerant of errors that occur during the character recognition process. The basic idea is to utilize the characteristics of recognition errors. One uses a confusion matrix to generate “equivalent” query strings that should match erroneously recognized text. The other one searches a “non-deterministic text” that contains multiple candidates for ambigous recognition results. Simulation experiments have shown that both methods can effectively combine character recognition with retrieval techniques.

A recursive analysis for form cell recognition

It is very difficult to analyze form structures because of breaks in lines and additional noises in the form image. This paper focuses on cell recognition in low quality form images. The recognition method has two features to achieve robustness in cell recognition. One is grid representation using several types of intersection and the terminal points of the frame lines. The other is the recursive modification of the representation. A new representation is created according to the determination of the breaks in the line and the hypothesized location of the missed intersections by using the previous representation. The modification is processed recursively until the representation has perfect consistency and all form cells are detected. In an experiment using 1565 form samples, all cells in 1538 samples (98.3% of 1565 samples) were recognized correctly by this method.

A new method of document structure extraction using generic layout knowledge

A method of document structure extraction using generic layout knowledge is described. With this method, it is possible to translate images of multimedia documents, i.e. documents that include pictures, graphics, and color information, to hypertext. Hypertext consists of decomposed elements linked with each other through some logical relationship. The principal components of the method are extraction of logical structure elements using a rectangular set operation and generation of hierarchical links of the logical structure between the extracted document elements. It is shown experimentally that the logical structure of a technical paper can be extracted.<<ETX>>

A high-speed rotation method for binary images based on coordinate operation of run data

A method of image rotation based on digital image data has been developed. This paper assumes the binary digital data, and proposes a method which is different from the traditional one based on pixel data. This method can execute a high-speed rotation of the binary image based on coordinate data for the start and the end of the run

Optimal techniques in OCR error correction for Japanese texts

This paper investigates three fundamental techniques in OCR error correction for Japanese texts using morphological analysis: (1) an optimal method for candidate word extraction from a candidate character lattice, (2) optimal word entries for Japanese verb inflection analysis, and (3) a new method of word matching cost calculation which is more suitable to be used with linguistic criteria. Comparative evaluation shows that the combination of these techniques requires 84% less computation, captures 2.6% more candidate words, reduces the chart parsing computation by 20%, and attains 25% higher error correction rate than a commonly used method.

Structure analysis method of graph image for document image retrieval

An automatic registration method for graph images which recognizes the structure of graphs is proposed. This method has the advantage of saving cost and time compared with the manual registration of types and labels of graphs as search indexes in electronic filing systems. The method is implemented in an advanced document image retrieval system, and the efficiency of the index registration is ascertained by experiment. The outline of this method is as follows: The types of graphs, the axes, and the labels of axes are initially detected. The types of graphs are classified into several categories and labels are attached to the axes. Next, the types are registered as indexes of the graphs, and the label images are recognized and converted into ASCII data, and also stored as indexes. An automaton model is used in the process of attaching the labels to analyze the components of the images. The automaton model refers to a knowledge-base of stored rules of layout structure. It then chooses appropriate labels for each axis.

A post-processing method for handwritten Kanji name recognition using Furigana information

Presents a new postprocessing method for recognizing handwritten Kanji names with Furigana, which shows the reading (pronunciation) of the Kanji in Kana characters, in real-time. Errors and rejects of Kanji OCR are automatically corrected to meaningful characters with the help of a Kanji-Furigana dictionary. Also, this method allows names to be written in any style. The method consists of Kanji word matching and Furigana verification. For Kanji word matching, Kanji name candidates are extracted. In Furigana verification, Kana candidate characters recognized by Kana OCR are compared with Furigana searched from the dictionary, using the Kanji name candidates. The correct name is selected from the results of Kanji and Kana word matching, plus the spacing between the first and second names. As result, 93.9% of the Kanji OCRs errors and rejects are corrected in real-time.<<ETX>>

Automatic template selection technique for the local pattern matching method

One of the effective means of detecting the position of the object by image processing is by local pattern matching which uses only a part of the image as the template. It is required for the implementation of this method to select a unique part of the image to be used as the template. This paper discusses the automatic selection of a template. As the first step, the causes of mismatching are discussed, and the distribution of matching error between the template and local patterns extracted from the image is analyzed. Based on the result, an evaluation function is defined which represents the uniqueness in the selection of the template. Auxiliary evaluation functions are also defined for pre-selection and reduction of computation time. These evaluation functions are both for determining the appropriateness of the template, and can be implemented on a simple dedicated circuit with practical processing time.

Defect detection method for stamped patterns utilizing random access parallel matching technique

In the fabrication processor of electronic components, the outlook inspection has been performed primarily by visual observation. The automation of this process is presently considered important. This paper describes automatic defect detection considering the stamped pattern on the surface of the electronic parts on the integrated circuit, etc. The only clue to the information on the kind of electronics part is the characters and symbols stamped on its surface. In this sense, the stamped pattern is considered as a part of the part quality. Some defects of the stamped pattern are the disappearance, lack of a part, and dirt. The feature of these defects are discussed first. The fixed-point sampling is considered which detects the existence of the character pattern at the observation point. The method is modified so that the decision can be made at the observation point along the stroke. Furthermore, a character defect detection system is proposed which can shift the observation point in horizontal and vertical directions so that the system can cope with the variation of the position of the character. A dedicated image processing device was developed to perform the defect detection with high speed. The architecture and the operation of the system are described. The system is advantageous in that the high-speed operation is realized by the pipeline control for a series of processings, from the parallel read-out of the local patterns in the image memory to the matching operation. Finally, the configuration of the experimental system is described and the results are shown for the automatic outlook inspection for the stamped pattern of the actual electronic part, thereby indicating the usefulness of the proposed system.

Consideration on automatic defect detection algorithm for stamped patterns in electronic parts

Central Research Laboratory, Hitachi, Ltd., Kokubunji, Japan 185 One of the most important problems in the quality inspection of printed characters, monitoring of printing devices, and evaluation of character recognition devices, is the automatic quality evaluation of the printed character for the character patterns stamped on the object surface or on a paper sheet. This paper considers the characters stamped on the surface of an electronic part such as transistor and integrated circuit, and reports on the result of comparison of defect detection methods for the automatic outlook inspection. First, properties of the stamped pattern, which is the object of inspection, are described, indicating the problems in detecting the defects such as lack, smudge and blur. Then three realizations of the defect detection are proposed: (1) a weighted matching is made with the reference pattern; (2) the reference pattern is divided spatially, and the matching is tried for the divided patterns; and (3) essential points are extracted from the core and background portions of the character, and a local pattern matching is tried. Finally, those defect detection methods are implemented on a computer, and an experiment was made for the actual stamped patterns on the transistor. The results for those methods are compared, and their effectiveness and the range of applications are indicated.