Back to results
US 5,347,594 AGranted
Method of image analysis
AssigneeGeneral Electric CGR S.A.
InventorMichel Grimaud
Life of this patent
Estimating the term…
Abstract
In order to analyze an image, a search is made, for all the minima of the image, for the dynamic of these minima. When the dynamic of these minima is small it corresponds finally to noise. If this measured dynamic is large, the minimum is regarded as being representative of a sufficiently significant irregularity in the image signal.
Claims20
A method of image analysis, which comprises the following steps:
- generating a signal of an image;
- locating the minima points of the image signal;
- determining a signal dynamic, for each of the minima points, which is dependent on the environment of the points neighboring the minimum point and which is constituted by the difference between the image signal value at the minimum point and a minimum maximorum image signal value for the neighboring points;
- extracting those minima points whose dynamic is greater than a threshold;
- and
- analyzing the image as a function of the list of points corresponding to the minima points extracted.
3 more elements
A method, comprising the steps of:
- (A) generating a signal representative of an image, said signal having minima values;
- (B) locating minima points corresponding to said minima values of said signal;
- (C) determining, for each of said minimum points, a signal dynamic which is dependent on the environment of points neighboring said minimum point and which is constituted by the difference between the image signal value at said minimum point and the minimum maximorum image signal value for said neighboring points;
- (D) comparing said dynamics of said minima points to a threshold;
- (E) extracting those minima points having dynamics which are greater than said threshold;
- (F) analyzing said image as a function of a list of points corresponding to the extracted ones of said minima points;
- and
- (G) displaying said image.
5 more elements
A system comprising:
- (A) means for generating a signal representative of an image, said signal having minima values;
- (B) means for locating minima points corresponding to said minima values of said signal;
- (C) means for determining, for each of said minima points, a signal dynamic which is dependent on the environment of points neighboring said minimum point and which is constituted by the difference between the image signal value at said minimum point and the minimum maximorum image signal value for said neighboring points;
- (D) means for comparing said dynamics of said minima points to a threshold;
- (E) means for extracting those minima points having dynamics the values of which are greater than said threshold;
- (F) means for analyzing said image as a function of a list of points corresponding to said extracted minima points;
- and
- (G) means for displaying said image.
5 more elements
A method of image analysis, which comprises the following steps:
- generating a signal of an image;
- locating the minima or maxima of the signal of the image;
- determining a signal dynamic, for each of the minima or maxima, which is dependent on the environment of the points corresponding to the corresponding minima or maxima;
- extracting those minima or maxima whose dynamic is greater than a threshold;
- and
- analyzing the image as a function of the list of points corresponding to the minima or maxima extracted;
- wherein
- each of the steps of determining a dynamic includes determining a dynamic for a minimum, and wherein each step of determining includes
- determining a neighborhood which is delimited by the points corresponding to a given change in contrast of the image, and
- measuring the difference of the signal of the image between the minimum and a point characteristic of the corresponding neighborhood;
- and further comprising the steps of
- for a given change in contrast, searching for image boundary points the value of whose image signal is less than that of the minimum, and
- retaining, as the neighborhood, intermediate points situated between the minimum and the boundary points.
10 more elements
A method of image analysis, which comprises the following steps:
- generating a signal of an image;
- locating the minima or maxima of the signal of the image;
- determining a signal dynamic, for each of the minima or maxima, which is dependent on the environment of the points corresponding to the corresponding minima or maxima;
- extracting those minima or maxima whose dynamic is greater than a threshold;
- and
- analyzing the image as a function of the list of points corresponding to the minima or maxima extracted;
- wherein
- each of the steps of determining a dynamic includes determining a dynamic for a minimum, and wherein each step of determining includes
- determining a neighborhood which is delimited by the points corresponding to a given change in contrast of the image, and
- measuring the difference of the signal of the image between the minimum and a point characteristic of the corresponding neighborhood;
- and further comprising the steps of
- choosing, as a point characteristic of the neighborhood, a point possessing, in a direction, a larger image signal than all the other points of the neighborhood in this same direction,
- searching for points which are similar to said point and which have the same properties in other directions, and
- choosing, among said similar points, the one point which has the smallest image signal as a characteristic point.
11 more elements
A method of image analysis, which comprises the following steps:
- generating a signal of an image;
- locating the minima or maxima of the signal of the image;
- determining a signal dynamic, for each of the minima or maxima, which is dependent on the environment of the points corresponding to the corresponding minima or maxima;
- extracting those minima or maxima whose dynamic is greater than a threshold;
- and
- analyzing the image as a function of the list of points corresponding to the minima or maxima extracted;
- wherein
- each of the steps of determining a dynamic includes determining a dynamic for a minimum, and wherein each step of determining includes
- determining a neighborhood which is delimited by the points corresponding to a given change in contrast of the image, and
- measuring the difference of the signal of the image between the minimum and a point characteristic of the corresponding neighborhood;
- and wherein the step of determining the dynamic includes
- (1) giving, to each point corresponding to a minimum of the image signal, a label representative of the respective minimum,
- (2) arraying all the image points corresponding to said minima in a hierarchical queue,
- (3) extracting said points corresponding to said minima from the hierarchical queue, beginning with the smallest value,
- (4) reentering the points that were the direct neighbors of the points extracted from the hierarchical queue at the time of extracting these points, if said direct neighbors are not already in the queue,
- (5) allocating, to the points entered into the hierarchical queue, a label identical to that of the point whose neighbors they are and whose extraction from the queue caused their entry, and
- (6) designating as a dynamic at said first minimum, the difference between the value of the image signal at said first minimum and the value of the image signal at a point having to be entered into the queue with the label of said first minimum as well as with the label of a second minimum but the value of whose image signal is smaller.
14 more elements
A method of image analysis, which comprises the following steps:
- generating a signal of an image;
- locating the minima or maxima of the signal of the image;
- determining a signal dynamic, for each of the minima or maxima, which is dependent on the environment of the points corresponding to the corresponding minima or maxima;
- extracting those minima or maxima whose dynamic is greater than a threshold;
- analyzing the image as a function of the list of points corresponding to the minima or maxima extracted;
- and
- contouring zones around the minima or maxima extracted according to a method of image processing with watershed.
4 more elements
A method, comprising the steps of:
- (A) generating a signal representative of an image, said signal having minima values;
- (B) locating points corresponding to said minima values of said signal;
- (C) determining, for each of said minima points, a signal dynamic which is dependent on the environment of points corresponding to the respective one of said minima values, said determining step including
- (i) determining a neighborhood which corresponds to said minimum andwhich is delimited by the points corresponding to a designated change in contrast of said image,
- (ii) measuring the difference, of said signals, between said minimum and a point characteristic of the corresponding neighborhood of said minimum,
- (iii) assigning, to a point corresponding to each of said minima, a label representative of the respective minimum,
- (iv) arranging all of said points corresponding to said minima in a hierarchical queue,
- (v) extracting said points corresponding to said minima from said hierarchical queue, beginning with the point having the smallest value,
- (vi) reentering those points, which were direct neighbors of said points extracted from said hierarchical queue at the time of extraction of said points, if said direct neighbors are not already in said hierarchical queue, and
- (vii) allocating, to the points entered in said hierarchical queue, a label identical to that of the points the neighbors of which and the extraction of which from said hierarchical queue, caused their entry;
- (D) comparing said dynamics of said minima points to a threshold;
- (E) extracting those minima points having dynamics which are greater than said threshold;
- (F) analyzing said image as a function of a list of points corresponding to the extracted ones of said minima points;
- and
- (G) displaying said image.
12 more elements
Dependencies are inferred from the claim text.