Fuzzy inference method and machine
First Claim
1. A fuzzy inference method for selecting a layout pattern on a display, performed by a computer, for inference based on fuzzy rules stated in the if-then form, comprising the steps of:
- (a) performing fuzzy inference based on the fuzzy rules to obtain inference results, (b) clustering the inference results into at least one cluster unit, after performing the fuzzy inference, (c) determining an integer value corresponding to a number of cluster units, (d) defuzzificating each of the cluster units based on the integer value determined in step (c) for producing a separate real value for each of the cluster units, and (e) selecting a layout pattern on the display as a function of the real values produced from the defuzzification.
1 Assignment
0 Petitions
Accused Products
Abstract
An image production method has steps of:
a detecting one or more descriptive regions exclusive of a background portion of a manuscript;
calculating an image characteristic value of said descriptive region;
selecting one from a plurality of previously defined layout patterns based on said determined image characteristic value, the layout pattern including information at least about a position of the descriptive region; and
producing an image by laying out an image of said manuscript descriptive region or images of an illustration and said manuscript descriptive region according to the decided layout pattern.
7 Citations
8 Claims
-
1. A fuzzy inference method for selecting a layout pattern on a display, performed by a computer, for inference based on fuzzy rules stated in the if-then form, comprising the steps of:
-
(a) performing fuzzy inference based on the fuzzy rules to obtain inference results, (b) clustering the inference results into at least one cluster unit, after performing the fuzzy inference, (c) determining an integer value corresponding to a number of cluster units, (d) defuzzificating each of the cluster units based on the integer value determined in step (c) for producing a separate real value for each of the cluster units, and (e) selecting a layout pattern on the display as a function of the real values produced from the defuzzification. - View Dependent Claims (2, 3, 4)
clustering the inference results into more than one cluster unit, after performing the fuzzy inference, determining the integer value as a value greater than one, and defuzzificating each of the cluster units to produce multiple real values based on the integer value determined to have a value greater than one.
-
-
5. A fuzzy inference computing machine for selecting a layout pattern on a display comprising:
-
fuzzy rule storage means of storing fuzzy rules;
fuzzy inference means of performing fuzzy inference based on the fuzzy rules stored in said fuzzy rule storage means to obtain inference results;
clustering means of clustering said inference results into at least one cluster unit and determining an integer value corresponding to a number of cluster units;
defuzzificating means of defuzzificating each of the cluster units based on the integer value to produce a separate real value for each of the cluster units; and
selection means of selecting a layout pattern on the display as a function of the real values produced from the defuzzificating means. - View Dependent Claims (6, 7)
the clustering means includes means of clustering the inference results into more than one cluster unit and determining the integer value as a value greater than one, and the defuzzificating means includes means of defuzzificating each of the cluster units to produce multiple real values based on the integer value having the value greater than one.
-
-
8. A fuzzy inference method performed by a computer for selecting a layout pattern on a display comprising the steps of:
-
(a) storing a plurality of objects;
(b) selecting attributes of a parameter under examination;
(c) storing a plurality of fuzzy rules for each attribute selected in step (b), in which each fuzzy rule provides a degree for favoring each object stored in step (a);
(d) calculating a characteristic value for at least one attribute selected in step (b);
(e) selecting sets of fuzzy rules from the plurality of fuzzy rules stored in step (c) based on the calculated characteristic value of step (d) to obtain inference results;
(f) clustering the inference results into at least one cluster unit after selecting the sets of fuzzy rules in step (e);
(g) determining an integer value corresponding to a number of cluster units;
(h) defuzzificating each of the cluster units obtained in step (f) based on the integer value determined in step (g) for producing a separate real value for each of the cluster units; and
(i) selecting an object belonging to a cluster unit having a real value defuzzificated in step (h) which is larger than other values defuzzificated in step (h) for each of the other cluster units, wherein the selected object is a layout pattern on the display.
-
Specification