Method and apparatus for remote tissue identification by statistical modeling and hypothesis testing of echo ultrasound signals
First Claim
1. A method for remote identification of a tissue using ultrasound, comprising the steps of:
- directing ultrasound energy into an unknown tissue;
measuring values of a characteristic of echoes of the ultrasound which are scattered from the unknown tissue;
accumulating a statistically significant sample of the measured values of the characteristic;
filtering the accumulated sample to determine, for each of a plurality of predetermined mathematical models, each of which is associated with a possible tissue type, a signal which is a measure of the likelihood that the accumulated sample of measured values was generated by said mathematical model;
applying the the filtered signals as inputs of a predetermined logical decision function which selects one of the models which most likely produced the sample; and
assigning the tissue type which is associated with the chosen model as the identity of the unknown tissue.
1 Assignment
0 Petitions
Accused Products
Abstract
Apparatus and methods for remote identification of tissue types model the scattering of ultrasound energy from living tissue as an autoregressive or autoregressive moving average random process. Autoregressive or autoregressive moving average models of candidate tissue types are generated from pulse-echo data that is known to come from that particular tissue type. Kalman prediction error filters are used for each candidate tissue type to generate estimates of the probability that an unknown pulse echo signal belongs to the class generated by that tissue type. Unknown pulse-echo signals are filtered in a specific Kalman filter to test the hypothesis that the unknown signal belongs to the class associated with that particular Kalman filter.
-
Citations
60 Claims
-
1. A method for remote identification of a tissue using ultrasound, comprising the steps of:
-
directing ultrasound energy into an unknown tissue; measuring values of a characteristic of echoes of the ultrasound which are scattered from the unknown tissue; accumulating a statistically significant sample of the measured values of the characteristic; filtering the accumulated sample to determine, for each of a plurality of predetermined mathematical models, each of which is associated with a possible tissue type, a signal which is a measure of the likelihood that the accumulated sample of measured values was generated by said mathematical model; applying the the filtered signals as inputs of a predetermined logical decision function which selects one of the models which most likely produced the sample; and assigning the tissue type which is associated with the chosen model as the identity of the unknown tissue. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33)
-
-
34. A method for manufacturing a filter which extracts, from a statistically significant sample of the values of signals which represent detected echoes of ultrasound which are scattered from unknown tissue, a signal which is a measure of the likelihood that the accumulated sample was generated by a mathematical model which characterizes a known tissue type, comprising the steps of:
-
directing ultrasound energy into a sample of the known tissue; detecting values characteristic of echoes of the ultrasound which are scattered from the known tissue; accumulating a statistically significant sample of the detected characteristic values; calculating, from the accumulated sample, autoregressive parameters which characterize an autoregressive process which models the sample values; and constructing a Kalman predictive filter which incorporates the determined autoregressive parameters. - View Dependent Claims (35, 36, 37, 38, 39)
-
-
40. Appartus for remote identification of a tissue comprising:
-
means for directing ultrasound energy into an unknown tissue; means for detecting a characteristic of echoes of the ultrasound which are scattered from the unknown tissue; means for accumulating a statistically significant sample of signals which represent values of the detected characteristic; a plurality of filter means, each of which function to filter the accumulated sample to extract a signal which is a measure of the likelihood that the accumulated sample was generated by an associated, predetermined mathematical model, wherein each of said mathematical models is associated with a possible tissue type; and means which apply the signals extracted by the filters, as inputs to a predetermined logical decision function to select the extracted signal which is associated with the model which most likely produced the sample and assign the tissue type associated with the model producing the chosen signal as the identity of the unknown tissue. - View Dependent Claims (41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60)
-
Specification