Method and apparatus for the acoustic emission monitoring detection, localization, and classification of metabolic bone disease

A noninvasive bone condition data acquisition system for sensing, receiving, processing, detecting, localizing and classifying transient, wideband acoustic emission signals emitted by rapid structural changes in the human musculo-skeletal system due to externally applied mechanical loads, comprising:

• at least one wideband transducer sensor for measuring tissue surface vibrations in skin;
• signal parameter estimator of at least one parameter from a digitized output of the wideband sensor to yield at least one information bearing attribute of the acoustic emission signal;
• feature extraction means for determining a feature vector from at least one information bearing attribute;
• and
• a signal processor for receiving said feature vector and comparing said feature vector to a library of known bone defects to yield a bone structure pattern assessment.

A method for noninvasive bone condition data acquisition for sensing, receiving, processing, detecting, localizing and classifying transient, wideband acoustic emission signals emitted by rapid structural changes in the human musculo-skeletal system due to externally applied mechanical loads comprising the steps of:

• estimating at least one parameter of a digitized output of said wideband sensor to yield at least one information bearing attribute of the acoustic emission signal resulting from mechanically stressed bone;
• extracting at least one feature vector with the at least one information bearing attribute as a component;
• and
• classifying the acoustic emission event using a signal processor by processing the at least one extracted feature relative to a library of such features recorded for various metabolic bone defects.

A method for noninvasive bone condition data acquisition for sensing, receiving, processing, detecting, localizing and classifying transient, wideband acoustic emission signals emitted by rapid structural changes in the human musculo-skeletal system due to externally applied mechanical loads, comprising the steps of:

• estimating at least one signal parameter at multiple observation points utilizing an array of wideband sensors while maintaining a set of relative times-of-arrival of the respective acoustic emission signals to determine to yield at least one information bearing attribute of the acoustic emission signal resulting from mechanically stressed bone while also recording a set of relative times of arrival of the respective emissions signals at the sensors;
• extracting a feature vector with at least one information bearing attribute as a component;
• classifying the acoustic emission event using a signal processor by processing the extracted feature relative to a library of such features recorded for various metabolic bone defects;
• and
• localizing the acoustic emission sources by spatial and temporal correlation processing the set of relative times-of-arrival.