Pattern inspection system using image correction scheme with object-sensitive automatic mode switchability
View Patent ↗An image correction device for use in a pattern inspection apparatus is disclosed, which has automatic adaptability to variations in density of a pattern image of a workpiece being tested. The device is operable to identify a two-dimensional (2D) linear predictive model from the pattern image of interest and determine the amount of eccentricity of a centroid position of this model. This amount is then used to switch between a corrected pattern image due to the 2D linear prediction modeling and a corrected image that is interpolated by bicubic interpolation techniques. A pattern inspection method using the image correction technique is also disclosed.
1. A device for correcting a pattern image by use of an inspection reference pattern image of a workpiece being tested and a pattern image under test, said device comprising:
an equation generation unit operative to generate by linear predictive modeling a set of simultaneous equations for the reference pattern image and the under-test pattern image;
a parameter generation unit operative to solve the equations to thereby obtain more than one model parameter;
a parameter centroid calculation unit operative to obtain a centroid position of the model parameter;
a sub-pixel interpolation computing unit operative to perform interpolation against the reference pattern image and the test pattern image to thereby generate a sub-pixel interpolated image;
a sub-pixel shift amount generation unit operative to use the sub-pixel interpolated image to obtain a sub-pixel shift amount for minimization of a deviation error in position between the reference pattern image and the test pattern image;
an error comparison unit operative to compare a difference between the centroid position of the model parameter and the sub-pixel shift amount to a threshold;
a corrected pattern image generation unit operative to use the model parameter to generate a corrected pattern image based on the linear predictive modeling; and
an interpolated image correction unit operative to generate an interpolated corrected image based on interpolation processing.
2. The device according to claim 1 , further comprising:
an image division unit for subdividing each of the reference pattern image and the test pattern image into a plurality of regions.
3. The device according to claim 1 , further comprising:
an image division unit for subdividing each of the reference pattern image and the test pattern image into a plurality of regions; and
a replacement correction unit for replacing part of the interpolated corrected image at a position corresponding to a divided corrected pattern image with the corrected pattern image.
4. The device according to claim 1 , further comprising:
a monitor unit for storing and monitoring a record history of at least one of the centroid position of the model parameter and the sub-pixel shift amount.
5. An apparatus for performing pattern inspection by using an inspection reference pattern image of a workpiece being tested and a pattern image under testing, said apparatus comprising:
an equation generation unit operative to generate by linear predictive modeling a set of simultaneous equations for the reference pattern image and the under-test pattern image;
a parameter generation unit operative to solve the equations to thereby obtain more than one model parameter;
a parameter centroid calculation unit operative to obtain a centroid position of the model parameter;
a sub-pixel interpolation computing unit operative to perform interpolation against the reference pattern image and the test pattern image to thereby generate a sub-pixel interpolated image;
a sub-pixel shift amount generation unit operative to use the sub-pixel interpolated image to obtain a sub-pixel shift amount for minimization of a deviation error in position between the reference pattern image and the test pattern image;
an error comparison unit operative to compare a difference between the centroid position of the model parameter and the sub-pixel shift amount to a threshold;
a corrected pattern image generation unit operative to use the model parameter to generate a corrected pattern image based on the linear predictive modeling;
an interpolated image correction unit operative to generate an interpolated corrected image based on interpolation processing; and
a pattern image comparison unit operative to compare the test pattern image to any one of the corrected pattern image and the interpolated corrected image.
6. A method for correcting a pattern image by use of an inspection reference pattern image of a workpiece being tested and a pattern image under test, said method comprising:
generating by linear predictive modeling a set of simultaneous equations for the reference pattern image and the under-test pattern image;
solving the equations to thereby obtain more than one model parameter;
obtaining a centroid position of the model parameter;
performing interpolation against the reference pattern image and the test pattern image to thereby generate a sub-pixel interpolated image;
using the sub-pixel interpolated image to obtain a sub-pixel shift amount for minimization of a deviation error in position between the reference pattern image and the test pattern image;
comparing a difference between the centroid position of the model parameter and the sub-pixel shift amount to a threshold;
when the difference is within the threshold, using the model parameter to generate a corrected pattern image based on the linear predictive modeling; and
when the difference is out of the threshold, generating an interpolated corrected image based on interpolation processing.
7. The method according to claim 6 , further comprising:
storing and monitoring a record history of at least one of the centroid position of the model parameter and the sub-pixel shift amount.
8. A method for correcting a pattern image by use of an inspection reference pattern image of a workpiece being tested and a pattern image under test, comprising:
subdividing each of the reference pattern image and the under-test pattern image into a plurality of regions;
generating by linear predictive modeling a set of simultaneous equations for the reference pattern image thus divided and the divided test pattern image divided;
solving the equations to obtain model parameters for the plurality of regions;
obtaining centroid positions of the model parameters with respect to said plurality of regions;
applying interpolation processing to the reference pattern image and the test pattern image in units of sub-pixels to thereby generate a sub-pixel interpolated image;
using the sub-pixel interpolated image to obtain a sub-pixel shift amount for minimization of a deviation error in position between the reference pattern image and the test pattern image;
comparing a difference between the centroid position of each model parameter and the sub-pixel shift amount to a threshold;
when the centroid position of every model parameter is within the threshold, combining together corrected pattern images as generated by the linear predictive modeling to thereby generate a single corrected pattern image; and
when the centroid position of at least one model parameter is out of the threshold, generating an interpolated corrected image based on interpolation processing.
9. A method for correcting a pattern image by use of an inspection reference pattern image of a workpiece being tested and a pattern image under test, comprising:
subdividing each of the reference pattern image and the under-test pattern image into a plurality of regions;
generating by linear predictive modeling a set of simultaneous equations for the reference pattern image thus divided and the divided test pattern image divided;
solving the equations to obtain model parameters for the plurality of regions;
obtaining a centroid position of model parameter for each said region;
applying interpolation processing to the reference pattern image and the test pattern image in units of sub-pixels to thereby generate a sub-pixel interpolated image;
using the sub-pixel interpolated image to obtain a sub-pixel shift amount for minimization of a deviation error in position between the reference pattern image and the test pattern image;
comparing a difference between the centroid position of each model parameter and the sub-pixel shift amount to a threshold;
when the centroid position of every model parameter is within the threshold, combining together corrected pattern images as generated by the linear predictive modeling to thereby generate a single corrected pattern image;
when the centroid position of at least one model parameter is out of the threshold, returning to said step of subdividing, thereby reducing a number of division of the reference pattern image and the test pattern image; and
in a case incapable of subdivision, when the centroid position of model parameter is out of the threshold, generating an interpolated corrected image based on interpolation processing.
10. A method of correcting a pattern image by use of an inspection reference pattern image of a workpiece being tested and a pattern image under test, comprising:
subdividing each of the reference pattern image and the under-test pattern image into a plurality of regions;
generating by linear predictive modeling a set of simultaneous equations for the reference pattern image thus divided and the divided test pattern image divided;
solving the equations to obtain model parameters for the plurality of regions;
obtaining centroid positions of the model parameters with respect to said plurality of regions;
applying interpolation processing to the reference pattern image and the test pattern image in units of sub-pixels to thereby generate a sub-pixel interpolated image;
using the sub-pixel interpolated image to obtain a sub-pixel shift amount for minimization of a deviation error in position between the reference pattern image and the test pattern image;
comparing a difference between the centroid position of each model parameter and the sub-pixel shift amount to a threshold;
when a difference between the centroid position of at least one model parameter and the sub-pixel shift amount is out of the threshold, generating an interpolated corrected image based on interpolation processing; and
performing image synthesis by cutting a region with the difference between the centroid position and the sub-pixel shift amount being out of the threshold and then pasting an interpolated corrected image corresponding such region to thereby generate a corrected pattern image.
11. A method of correcting a pattern image by use of an inspection reference pattern image of a workpiece being tested and a pattern image under test, comprising:
subdividing each of the reference pattern image and the under-test pattern image into a plurality of regions;
generating by linear predictive modeling a set of simultaneous equations for the reference pattern image thus divided and the divided test pattern image divided;
solving the equations to obtain model parameters for the plurality of regions;
obtaining centroid positions of the model parameters with respect to said plurality of regions;
applying interpolation processing to the reference pattern image and the test pattern image in units of sub-pixels to thereby generate a sub-pixel interpolated image;
using the sub-pixel interpolated image to obtain a sub-pixel shift amount for minimization of a deviation error in position between the reference pattern image and the test pattern image;
comparing a difference between the centroid position of each model parameter and the sub-pixel shift amount to a threshold; and
when a difference between the centroid position of at least one model parameter and the sub-pixel shift amount is out of the threshold, performing image synthesis by cutting a corrected pattern image region thereof and then pasting in the cut region a corrected pattern image with a finally obtained difference being within the threshold to thereby generate a corrected pattern image.
12. A method for performing pattern inspection by using an inspection reference pattern image of a workpiece being tested and a pattern image under testing, comprising:
generating by linear predictive modeling a set of simultaneous equations for the reference pattern image and the under-test pattern image;
solving the equations to thereby obtain more than one model parameter;
obtaining a centroid position of the model parameter;
performing interpolation against the reference pattern image and the test pattern image to thereby generate a sub-pixel interpolated image;
using the sub-pixel interpolated image to obtain a sub-pixel shift amount for minimization of a deviation error in position between the reference pattern image and the test pattern image;
comparing a difference between the centroid position of the model parameter and the sub-pixel shift amount to a threshold;
when the difference is within the threshold, using the model parameter to generate a corrected pattern image based on the linear predictive modeling;
when the difference is out of the threshold, generating an interpolated corrected image based on interpolation processing; and
comparing the test pattern image to any one of the corrected pattern image and the interpolated corrected image.