| US 7,394,927 B2 | ||
| Iterative method of determining a spatial distribution of values of a property | ||
| Roland Proksa, Hamburg (Germany); and Thomas Köhler, Norderstedt (Germany) | ||
| Assigned to Koninklijke Philips Electronics N.V., Eindhoven (Netherlands) | ||
| Appl. No. 10/555,576 PCT Filed Apr. 26, 2004, PCT No. PCT/IB2004/050507 § 371(c)(1), (2), (4) Date Sep. 12, 2006, PCT Pub. No. WO2004/100070, PCT Pub. Date Nov. 18, 2004. |
||
| Claims priority of application No. 03101249 (EP), filed on May 06, 2003. | ||
| Prior Publication US 2007/0003132 A1, Jan. 04, 2007 | ||
| Int. Cl. G06K 9/00 (2006.01) | ||
| U.S. Cl. 382—154 [382/128; 382/255; 382/131; 324/307; 324/309; 324/318; 324/319] | 10 Claims |
| 1. An iterative method of determining a spatial distribution of values (μj) of a property of an object, and particularly values of its absorption, in an examination region, starting from measured
values (pl) that are first acquired with a measuring device, and particularly with a computer tomograph, wherein the measured values
(pl) can each be represented as a sum of values (μj) of a property that have each been multiplied by a proportional factor (Alj), which measured values are divided into groups, and wherein the proportional factor (Alj) is a measure of the proportion that a value (μj) of the property forms of the measured value (pl), and each value (μj) of the property is approached by one iteration value (μjk) at a time by setting each iteration value (μjk) to a starting value and performing the following steps at least once on each group of measured values (pl) until a stop criterion is satisfied:
a) generating a reference measured value for each measured value (pl) in the given group, by multiplying each iteration value (μjk) whose value (μj) of the property forms a proportion of the given measured value by a proportional factor (Alj) and by adding up the iteration values that are multiplied by the given proportional factor (Alj),
b) generating a measured value difference for each measured value (pl) in the given group by forming the difference between the given measured value (pl) and the corresponding reference measured value,
c) for each measured value (p2) selecting a corresponding reliability parameter (λl) and multiplying each measured value difference by the corresponding reliability parameter (λl), the greater the reliability of that measured value (pl), the larger the corresponding reliability parameter (λl),
d) dividing each measured value difference that has been multiplied by the reliability parameter (λl) by a normalizing factor that is dependent on the proportional factors of those values of the property which form a proportion
of that measured value which belongs to the measured value difference,
e) updating the distribution by backward projection of the measured value differences obtained in step d) into the examination
region.
|