| 摘要 |
An image region extraction capable of extracting an image of a significant moving object in a dynamic image accurately and easily at high speed. The operator is required to manually enter the rough boundaries for the significant region in the first two fields of the dynamic image only, and the subsequent extraction operation for each subsequent field of the dynamic image is carried out automatically according to the conjectured rough boundary obtained on a basis of the extraction regions for the two fields immediately preceding that field. The extraction region for the significant region is obtained from the rough boundary on a basis of an original image of a respective field of the dynamic image, and the extracted image of the significant region in each field of the dynamic image is obtained by multiplying the original image of each field and the extraction region for the significant region in each field. <IMAGE> |
