Method and apparatus for predicting values of a varying periodic phenomenon

• US 5,051,903 A
• Filed: 08/14/1989
• Issued: 09/24/1991
• Est. Priority Date: 08/14/1989
• Status: Expired due to Term

• Current Assignee: General Electric Company
• Original Assignee: General Electric Company
• Inventors: Pelc, Norbert J., Flax, Stephen W.
• Primary Examiner(s): Smith, Jerry
• Assistant Examiner(s): Lo, Allen M.
• Application Number:
  US07/393,376
• Time in Patent Office: 771 Days
• Field of Search: 364/413.13, 324/309
• US Class Current: 324/309
• CPC Class Codes: G01R 33/5673 Gating or triggering based on a physiological signal other than an MR signal, e.g. ECG gating or motion monitoring using optical systems for monitoring the motion of a fiducial marker

Abstract

An apparatus for reducing image artifacts in NMR imaging matches elements of a set to values of a substantially periodic function so that the values exhibit a predetermined relationship to the elements. The matching is performed by evaluating the relative probability of the values of the substantially periodic function from the samples in the growing database of the values and assigning the values to the elements by using the evaluated relative probability, so as to maximize the probability that subsequent valves may be assigned the remaining elements according the predetermined relationship. A method for correcting the matching of values with elements freezes the database and acquires additional values which are substituted for earlier matched values based on the frozen database.

Claims

1. A method of associating a sampled value y(t_i) of a substantially periodic signal y(t), with an element G_y (i) in a set G_y, such that the sampled value y(t_i) has a predetermined relationship to the order of the element G_y within the set, comprising the steps of:

(a) acquiring a database of N_t samples of the substantially periodic signal y(t);

(b) growing the database by adding the sampled value y(t_i) to the database without removing another sample from the database;

(c) evaluating the relative probability of the values of the substantially periodic function y(t) from the samples in the growing database; and

(d) assigning the value y(t_i) to an element G_y (i), using the evaluated relative probability, so as to maximize the probability that subsequent values of y(t) may be assigned the remaining elements of set G_y according the predetermined relationship.

2. The method of claim 1 wherein step (a) further comprises rolling the database by storing new samples of the substantially periodic signal y(t) in the data base and discarding the oldest samples stored in the database so as to maintain N_t samples.

3. The method of claim 1 wherein steps (c) and (d) are performed by:

assigning a reference phase φ

_ref (i) to each element G_y (i) of the set G_y such that the relationship between the value of the reference phase φ

_ref (i) and the order of the element G_y (i) in the set G_y is the same as that desired between the value of the substantially periodic signal y(t_i) and the order of the element G_y (i);

producing a current phase φ

(y_i) from the current value of the periodic signal y(t_i), by using the database, so that the probability density of the current phase values φ

(y_i) produced is substantially more uniform than the probability density of the current values of the periodic signal y(t_i); and

associating the current value of the periodic signal y(t_i) with an element of the set G_y (i) according to the differences between the current phase φ

(y_i) and the closest available reference phase φ

_ref (i).

4. The method of claim 1 wherein the values of a periodic signal y(t_i) are values of physiological motion of an imaged body during an NMR imaging scan and wherein the elements of the ordered set G_y (i) are magnitudes of a phase encoding gradient and wherein the relationship between the values of the periodic function and the element'"'"'s order is selected to reduce motion artifacts resulting from the physiological motion.

5. The method of claim 1 wherein the values of the periodic function y(t) are scaled and offset by automatic gain control (AGC) circuitry and automatic offset control (AOC) circuitry prior to being stored in the data base and wherein the scaling and offset values are fixed prior to step (b).

6. The method of claim 1 wherein the values of the periodic function y(t) are scaled and offset by automatic gain control (AGC) circuitry and automatic offset control (AOC) circuitry prior to being stored in the data base and wherein the scaling and offset values are fixed prior to step (b).

7. A method of associating a value y(t_i) of a substantially periodic signal y(t), with an element G_y (i) in a set G_y, such that the value y(t_i) has a predetermined relationship to the order of the element G_y (i) within the set, comprising the steps of:

(a) recording in a database N_t samples of the substantially periodic signal y(t);

(b) freezing the database by neither adding nor removing samples;

(c) evaluating the relative probability of the values of the substantially periodic function y(t) from the samples in the frozen database; and

(d) assigning the value y(t_i) to an element G_y (i), using the evaluated relative probability, so as to maximize the probability that subsequent values of y(t) may be assigned the remaining elements of set G_y according the predetermined relationship.

8. The method of claim 7 wherein step (a) further comprises rolling the database by storing new samples of the substantially periodic signal y(t) in the data base and discarding the oldest samples stored in the database so as to maintain N_t samples.

9. The method of claim 7 wherein step (d) is performed by:

assigning a reference phase φ

_ref (i) to each element G_y (i) of the set G_y such that the relationship between the value of the reference phase φ

_ref (i) and the order of the element G_y (i) in the set G_y is the same as that desired between the value of the substantially periodic signal y(t_i) and the order of the element G_y (i);

producing a current phase φ

(y_i) from the current value of the periodic signal y(t_i), by using the database, so that the probability density of the current phase values φ

(y_i) produced is substantially more uniform than the probability density of the current values of the periodic signal y(t_i); and

associating the current value of the periodic signal y(t_i) with an element of the set G_y (i) according to the differences between the current phase φ

(y_i) and the closest available reference phase φ

_ref (i).

10. The method of claim 9 wherein step (a) is performed by growing the database by adding the sample y(t_i) to the data base without removing an older sample from the database.

11. The method of claim 9 wherein step (a) further is performed by rolling the database by storing new samples of the substantially periodic signal y(t) in the data base and discarding the oldest samples stored in the database so as to maintain N_t samples.

12. The method of claim 7 wherein the values of a periodic signal y(t_i) are values of physiological motion of an imaged body during an NMR imaging scan and wherein the elements of the ordered set G_y (i) are magnitudes of a phase encoding gradient and wherein the relationship between the values of the periodic function and the element'"'"'s order is selected to reduce motion artifacts resulting from the physiological motion.

13. A method of associating a values of a substantially periodic signal y(t), with the elements of set G_y, such that the values have a predetermined relationship to the order of the elements within the set, comprising the steps of:

(a) acquiring a database of values of the substantially periodic signal y(t);

(b) evaluating the relative probability of the values of the substantially periodic function y(t) from the samples in the database;

(c) assigning a value y(t_i) of the substantially periodic signal y(t) to an element G_y (i), by using the evaluated relative probability, so as to maximize the probability that subsequent values of y(t) may be assigned the remaining elements of set G_y according the predetermined relationship;

(d) freezing the database by ceasing to record values of the substantially periodic signal y(t) subsequent to the time when each element G_y (i) has been associated with a value y(t_i);

(e) sampling the substantially periodic waveform y(t) to obtain an overscan value y_s (t_n); and

(f) replacing a previously assigned value y(t_i) with the overscan value y_s (t_n) so that the relationship between the assigned values of the substantially periodic signal y(t) and the elements of set G_y is made closer to the predetermined relationship.