Dainippon Screen Mfg. Co., Ltd.

Junichi Shiomi

Atsushi Imamura

Hiroshi Sano

Eiji Nishihara

Takashi Sakamoto

Hiroki Fujimoto

Takahide Hirawa

The present invention provides a technique to identify layout of surface regions of an object by image area segmentation. A plurality of images is acquired by imaging an object under a plurality of mutually different illumination conditions. By performing area segmentation on this plurality of images, the layouts of the plurality of image areas which is different from the layout of the object surface regions is obtained. Then, the layout of the surface regions on the object is identified based on the layouts of the plurality of image areas.

Surface region identification by image area segmentation

A halftone dot image HI composed of halftone dots is produced by halftone dot meshing of an object image OI. A smoothed object image SOI and a smoothed halftone dot image SHI are then produced by smoothing the object image and the halftone dot image. An interference moire image IM is produced by obtaining a difference between the smoothed object image SOI and the smoothed halftone dot image SHI.

Evaluation of dot interference moir&#xe9;

A color image of a circuit board is divided into respective color regions. Plural representative colors are set, and angle indices and distance indices are calculated for each pixel color in the color image in a predetermined color space. The angle indices for a particular pixel color represent angles between an individual color vector representing the particular pixel color and plural representative color vectors of the plural representative colors. The distance indices for a particular pixel color represent distances between the particular pixel color and the plural representative colors. Composite distance indices are then calculated for each pixel color in the color image, based on the distance indices and the angle indices. Each pixel in the color image is classified into plural representative color regions associated with the plural representative colors according to the composite distance indices, thereby dividing the image region of the color image into the plural representative color regions.

Region segmentation of color image

A two-dimensional image of a substrate surface targeted for polishing is periodically picked up, and the image is analyzed to obtain an entropy H, Hof the two-dimensional image. An end point of polishing is then determined according to the entropy H, H. Alternatively, other image characteristic value such as a difference statistic of the image may be employed instead of the entropy H, H

Detection of an end point of polishing a substrate

Each recording pixel is divided into M small divisions where M is an integer of not less than 2. A threshold pattern is provided for the small divisions where a threshold value is assigned to each of the M small divisions corresponding to each recording pixel. The threshold values assigned to the M small divisions are compared with a multi-tone image signal to produce M pieces of binary comparison results. An image recording signal is generated as a function of a sum n of the M pieces of binary comparison results.

Method and apparatus for recording halftone image utilizing tone reproduction at each pixel

Threshold values are generated in order to disperse isolated dots in an irregular manner when an level of an image signal is between a first signal level and a second signal level. The first signal level is a level at which one isolated dot consisting of one pixel is formed in each one-dot area including one halftone dot, and the second signal level is a level at which no isolated dot consisting of one pixel is formed in each one-dot area. The threshold values are then compared with the image signal, to generate a halftone dot signal representing halftone dots.

Method and apparatus for generating halftone dots for color printing

An exposure signal representing a halftone image is generated by executing tone correction on at least either of an image signal and n pieces of dot pattern signals, and by comparing the image signal with the n pieces of dot pattern signals after the tone correction. The tone correction characteristics are set to compensate for a difference in dot gain characteristics between dot types. The halftone image is accordingly reproduced with the compensated dot gain characteristics.

Growing community of inventors

Kyoto, Japan

Junichi Shiomi