摘要 |
<P>PROBLEM TO BE SOLVED: To provide a digital image data analysis method with which three-dimensional (3D) image display and automatic diagnosis can be performed from digital image data by making diagnosis or determination, by analyzing the digital image data, in a specified portion or region within an image although it is performed conventionally by human visual sensation, and a diagnosis system employing the same. <P>SOLUTION: As a method for separating a required region and another region while utilizing pixel values of digital image data, by setting any arbitrary conversion rate to the pixel values, the pixel values are made different, thereby making the required region and the other region differential and independent. For a boundary value between the required region and the other region, fuzzy learning vector quantization (FLVQ) is used for the discrimination, separation precision is ensured, the required region is extracted by labeling, and conversion into a 3D image is performed by superimposing image data of the extracted region, thereby displaying a stereoscopic shape. <P>COPYRIGHT: (C)2012,JPO&INPIT |