Image inspection methods and apparatus
AssigneePhilip Morris Incorporated
InventorDaniel Ennis
Life of this patent
Estimating the term…
Abstract
An image is tested to determine whether it is substantially like a plurality of sample images by computing statistical information about the sample images and using that statistical information to analyze the image being tested. The analysis is such that the statistical parameters are standardized so that available tables of the central chi-square distribution function can be used, thereby simplifying the necessary calculations. If desired, images which have been determined by the method of this invention to be good can be used to refine the statistical information used in analyzing subsequent images. When an image is identified as unacceptable by the method of the invention, the data for that image can be decomposed in order to identify the parts of the image which are causing the image to be unacceptable.
Claims14
The method of determining whether a test image is substantially like any of a plurality of sample images comprising the steps of:
- using said computer processor means to subdivide each of said sample images into a plurality of segments;
- using said computer processor means to associate with each segment of each sample image a value representative of the image information in the segment to produce a vector of said values for each sample image, each vector having a number of elements equal to the number of segments in each sample image;
- using said computer processor means to calculate a variance-covariance matrix of said vectors;
- using said computer processor means to calulate a mean vector of said vectors;
- using said computer processor means to calculate a Cholesky decomposition matrix of said variance-covariance matrix such that said Cholesky decomposition matrix times the transpose of said Cholesky decomposition matrix equals said variance-covariance matrix;
- using a camera to capture a test signal representative of said test image;
- using computer processor means to form said test image from of said test signal;
- using said computer processor means to convert said test image into gray-scale pixel density values;
- using said computer processor means to subdivide said test image into a plurality of segments similar to the segments used for said sample images;
- using said computer processor means to associate with each segment of said test image a value representative of the image information in that segment to produce an image vector of said values, said image vector having a number of elements equal to the number of segments in said test image;
- using said computer processor means to calculate a product vector equal to the product of the inverse of said Cholesky decomposition matrix times the difference between said image vector and said means vector;
- using computer processor means to calculate the dot product of said product vector with itself;
- using computer processor means to compare said dot product to a predetermined value;
- and
- using said computer processor means to identify said test image as substantially like a random sample of said sample images only if said dot product has a predetermined relationship to said predetermined value.
Apparatus for determining whether a test image is substantially like any of a plurality of sample images comprising:
- computer processor means for similarly subdividing each of said sample images into a plurality of segments;
- computer processor means for associating with each segment of each sample image a value representative of the image information in that segment to produce a vector of said values for each sample image, each vector having a number of elements equal to the number of segments in each sample image;
- computer processor means for computing a variance-covariance matrix of said vectors;
- computer processor means for computing a mean vector of said vectors;
- computer processor means for computing a Cholesky decomposition matrix of said variance-covariance matrix such that said Cholesky decomposition matrix times the transpose of said Cholesky decomposition matrix equals said variance-covariance matrix;
- camera means for capturing a test signal representative of said test image;
- computer processor means for forming said test image from said test signal;
- computer processor means for converting said test image into gray-scale pixel density values;
- computer processor means for subdividing said test image into a plurality of segments similar to the segments used for said sample images;
- computer processor means for associating with each segment of said test image a value representative of the image information in that segment to produce an image vector of said values, said image vector having a number of elements equal to the number of segments in said test image;
- computer processor means for computing a product vector equal to the product of the inverse of said Cholesky decomposition matrix times the difference between said image vector and said mean vector;
- computer processor means for computing the dot product of said product vector with itself;
- computer processor means for comparing said dot product to a predetermined value;
- and
- computer processor means for identifying said test image as substantially like a random sample of said sample images only if said dot product has a predetermined relationship to said predetermined value.
Dependencies are inferred from the claim text.