This paper proposes a system for the inspection of character printing quality. This system extracts printed character images from background figures using color image transformation and detects defects both of the printing density and of the printed character shape. To measure defects of printing density, the system extracts pixels whose gray level is between the levels of characters and the background area and measures distances to the nearest background pixels. Matching with the maximum/minimum outline limits of characters and comparing their topological features with the standard data are adopted to the detect defects of the shape. Experiments data compared with human inspection have also been described.

View full abstract

This paper proposes a multi-font Japanese character recognition method using direction frequency distribution among pixels that construct character-lines (Directional Feature) and spectra around a character (Peripheral Feature). These two features seem to be a quality that they are stable against the distortion of character forms, changes of shift and so on. Besides, we discuss a scheme for placing small areas used in directional feature extraction and that for placing partitions used in peripheral feature extraction. On the basis of the above discussion, we propose a feature extraction method that would be useful for more precision recognition. Finally, some experiments were done for 10215 Japanese characters (Hiragana. Katakana and Chinese character) of five fonts(Mincho, Socho, Gothic, Kyokasho and Square style). As a result, the recognition rate overheads 99.22%.

View full abstract

A devlopment of CT and MR scanners enables us to obtain the human internal images. However, it is difficult to extract some volumes of various brain parts from the images obtained from them, because the images are 2D slice images. Therefore it is necessary to do the medical image processing as a technique that makes us understand the obtained images. This paper proposes an automatic extraction method for cerebral region based on region growing. This method enables us to obtain the volumes of various bra,in parts. Moreover, it shows a method to do 3D display of the extracted shape.

View full abstract

We build the image processing expert, system which uses the process of the processing. Without the generation of the procedure of the processing with the production system, it chooses a procedure from datas of the process of the processing and reuses it. In the process of the processing, the redundant information of repeated trial and error is contained. However, as the knowledge which was met with, making the record of the processing knowledge and analyzing it,, it, stands in the duty to the beginner. And then, as applying this system to general image processing, good evaluation was gotten to the reuse of the process of the processing.

View full abstract

Image database systems require powerful navigation tools to assist users who have only a vague idea, and to help novice users. To develop such an advanced and flexible image retrieval mechanism, it is important to take account of user's intention that is shown where they look at in data attributes. This paper presents a new approach for image retrieval systems based on user's viewpoints. And we have made a prototype by using a large scale knowledge base what is capable of processing natural languages. Some experimental results on the image database which stores about 1000 images are also presented.

View full abstract

The algorithm that we have formerly described always rewrites labels on the immediately preceding row, using a label LRAM(LRAM(Z)) instead of label Z in a labeling logical mask, where LRAM is one-dimensional table for label linkage information. Moreover, it also unifies and updates the label's linkage information in the table LRAM by the row-by-row unification processing. However, this row-by-row unification causes a problem on how to secure such spared time to do the unification. This paper presents a new hardware-oriented algorithm without using the unification. When raster-scanning, this algorithm associates the leftmost point label on a pattern with the minimum label of the connected region, at each row. And at the next row, It rewrites the leftmost point label into the minimum label and propagates the label from left to right by rewriting other connected labels and the content of LRAM. Then, it brings about the same results as the row-by-row unification. This paper also shows that the hardware realization is possible if the LRAM is three times accessible within time of one pixel processing.

View full abstract

The aim of this paper is to pressent, for the reader, the basic result that is oftained by lilerture[4] . This result can be briefly formulated and prove as follows : the symmetrical figures are constracted by Origami Transform 2 or 3 . Yatsuko- assembly, set of Yatsuko - assembly and parallel-assembly, set of parallel - assembly are newly difined in Sec 3 . The constructed symmetrical figures are concretely indicated by 1 ( set of point, in Sec 5), by 2 ( set of line, in Sec 6), by 3 (general figures, in Sec 7).

View full abstract