Back to results
US 6,058,352 AGranted
Accurate tissue injury assessment using hybrid neural network analysis
AssigneePhysical Optics Corporation
InventorsTaiwei Lu, Robert A. Lieberman
Life of this patent
Estimating the term…
Abstract
Systems and methods using a neural network based portable absorption spectrometer system for real-time automatic evaluation of tissue injury are described. An apparatus includes an electromagnetic signal generator; an optical fiber connected to the electromagnetic signal generator; a fiber optic probe connected to the optical fiber; a broad band spectrometer connected to the fiber optic probe; and a hybrid neural network connected to the broad band spectrometer. The hybrid neural network includes a principle component analyzer of broad band spectral data obtained from said broad band spectrometer.
Claims13
An apparatus for assessing an injury to tissue, comprising:
- an electromagnetic signal generator;
- an optical fiber connected to said electromagnetic signal generator;
- a fiber optic probe connected to said optical fiber;
- a broad band spectrometer connected to said fiber optic probe;
- and
- a hybrid neural network connected to said broad band spectrometer, said hybrid neural network having a plurality of output neurons, each of said plurality of output neurons corresponding to a tissue injury category;
- and
- wherein said hybrid neural network includes a principle component analyzer of broad band spectral data obtained from said broad band spectrometer;
- and
- wherein said hybrid neural network produces a set of results at said plurality of output neurons, said set of results indicating the nature of said injury to said tissue.
7 more elements
A method of assessing an injury to tissue, comprising:
- (A) providing
- (1) an electromagnetic signal generator;
- (2) an optical fiber connected to said electromagnetic signal generator;
- (3) a fiber optic probe connected to said optical fiber;
- (4) a broad band spectrometer connected to said fiber optic probe;
- (5) a hybrid neural network connected to said broad band spectrometer, said hybrid neural network having a plurality of output neurons, each of said plurality of output neurons corresponding to a tissue injury category;
- and
- (6) an output device connected to said hybrid neural network;
- (B) positioning said fiber optic probe proximal an object to be analyzed;
- (C) transmitting a first electromagnetic signal to said object from said electromagnetic signal generator;
- (D) conveying a second electromagnetic signal from said object to said broad band spectrometer;
- (E) transforming said second electromagnetic signal from said object into a broad band spectrum with said broad band spectrometer;
- (F) transmitting said broad band spectrum from said broad band spectrometer to said hybrid neural network;
- (G) processing said broad band spectrum with said hybrid neural network to obtain an autoscaling of said broad band spectrum;
- (H) processing said autoscaling of said broad band spectrum with said hybrid neural network to obtain a principal component analysis of said broad band spectrum by extracting a set of orthogonal feature vectors to represent said broad band spectrum;
- (I) classifying said set of orthogonal feature vectors with said hybrid neural network to produce a set of results at said plurality of output neurons, said set of results indicating the nature of said injury to said tissue;
- and
- (J) displaying a representation of said set of results using said output device.
15 more elements
An apparatus for assessing an injury to tissue, comprising:
- an electromagnetic signal generator;
- an optical fiber connected to said electromagnetic signal generator;
- a fiber optic probe connected to said optical fiber;
- a broad band spectrometer connected to said fiber optic probe;
- and
- a hybrid neural network connected to said broad band spectrometer, said hybrid neural network including:
- a first means for autoscaling broad band spectral data obtained from said broad band spectrometer;
- a means for principal component analysis of broad band spectral data obtained from said broad band spectrometer, said means for principal component analysis including means for performing singular value decomposition analysis;
- a second means for autoscaling data from said means for principal component analysis;
- a neural network classifier;
- and
- a plurality of output neurons, said plurality of output neurons being disposed at the output of the neural network classifier, each of said plurality of output neurons corresponding to a tissue injury category;
- and
- wherein said hybrid neural network produces a set of results at said plurality of output neurons, said set of results indicating the nature of said injury to said tissue.
11 more elements
Dependencies are inferred from the claim text.