Image detecting apparatus

Claim 2

An image area extracting method for extracting a desired area from an image, comprising the steps of: dividing rough boundary area image corresponding to said image, in which boundary area between said desired area and areas other than said desired area is roughly assigned, into a plurality of small areas; detecting the desired area and areas other than said desired area from said each small area by executing area extracting means to each of said small areas; and forming a boundary area mask image of said roughly assigned boundary area between said desired area and areas other than said desired area depending upon said detection result and extracting said desired area based on said boundary mask image, wherein said image area extracting processing comprises extracting the image area by utilizing an axis selection edge detecting process wherein edges of the desired area are detected by selecting an edge detection axis direction on the image and alternatively by selecting an edge detection axis direction on color space.

Claim 3

The image area extracting method according to claim 1, wherein a parameter of one of said each area extracting means which is executed to said each small area of said detecting step can be adjusted, and the parameter adjusted for each small area is applied.

Claim 4

The image area extracting method according to claim 2, wherein at least one of said area extracting means of said detecting step is the processing of approximating color distribution obtained by distributing color information of designated pixel on color space with an ellipsoid and dividing the image area by dividing said color space using said ellipsoid.

Claim 5

The image area extracting method according to claim 1, wherein when there are at least two area extracting means which are executed to said each small area of said detecting step, an optimum area extracting means is selected for said each small area to execute the selected area extracting means for each small area.

Claim 6

The image area extracting method according to claim 4, wherein a parameter of one of said each area extracting means which is executed to said each small area of said detecting step can be adjusted, and the parameter adjusted for each small area is applied.

Claim 7

The image area extracting method according to claim 5, wherein at least one of said area extracting means of said detecting step is the processing of approximating color distribution obtained by distributing color information of designated pixel on color space with an ellipsoid and dividing the image area by dividing said color space using said ellipsoid.

Claim 8

An image area extracting method for extracting a desired area from an image, comprising the steps of: dividing rough boundary area image corresponding to said image, in which boundary area between said desired area and areas other than said desired area is roughly assigned, into a plurality of small areas; detecting the desired area and areas other than said desired area from said each small area by executing area extracting means to each of said small areas; forming a boundary area mask image of said roughly assigned boundary area between said desired area and areas other than said desired area depending upon said detection result; forming a non-boundary area mask image of said areas other than said roughly assigned boundary area; extracting said desired area based on said boundary mask image; and synthesizing the boundary area mask image and the non-boundary mask image to form a synthesized mask image for assigning the desired area, wherein said image area extracting processing comprises extracting the image area by utilizing an axis selection edge detecting process wherein edges of the desired area are detected by selecting an edge detection axis direction on the image and alternatively by selecting an edge detection axis direction on color space.

Claim 9

An image area extracting method for extracting a desired area from an image, comprising the steps of: dividing rough boundary area image corresponding to said image, in which boundary area between said desired area and areas other than said desired area is roughly assigned, into a plurality of small areas; detecting the desired area and areas other than said desired area from said each small area by executing area extracting means to each of said small areas; forming a boundary area mask image of said roughly assigned boundary area between said desired area and areas other than said desired area depending upon said detection result; forming a non-boundary area mask image of said areas other than said roughly assigned boundary area; extracting said desired area based on said boundary mask image; and synthesizing the boundary area mask image and the non-boundary mask image to form a synthesized mask image for assigning the desired area, wherein said image area extracting processing comprises extracting the image area by utilizing an axis selection edge detecting process wherein edges of the desired area are detected by selecting an edge detection axis direction on the image and alternatively by selecting an edge detection axis direction on color space.