System and method of using Multi Pattern Viterbi Algorithm for joint decoding of multiple patterns

1. A method, comprising:

receiving a set of signal observations for a signal repetition, wherein the set of signal observations is associated with a dimension of a K-dimensional time grid having a plurality of time-indexed points;

calculating a state cost metric for a state in a plurality of states of a statistical model, for a given time-indexed point, wherein the state cost metric calculation provides a most-likely predecessor state and a corresponding most-likely predecessor time-indexed point;

determining a sequence of states using the state cost metric calculation;

determining a corresponding cumulative probability measure for the statistical model;

calculating a cost metric associated with a possible prior state at a possible predecessor time-indexed point;

selecting a lowest cost metric for the state to determine the state cost metric;

observing dimensions that are incremented when moving from a given predecessor time-index point to a given one of the plurality of time-indexed points; and

basing the state cost metric on at least the observing.

2. The method of claim 1 , further comprising:

repeating the calculating the state cost metric for the state for the plurality of time-indexed points and determining a most likely sequence and corresponding cumulative probability measure for a plurality of statistical models; and

identifying a most likely statistical model based on the corresponding cumulative probability measures for the plurality of statistical models.

3. The method of claim 1 , wherein the determining the sequence of states comprises identifying a lowest state cost metric at a final state at a terminal time-indexed point.

4. A method, comprising:

receiving a set of signal observations for a signal repetition, wherein the set of signal observations is associated with a dimension of a K-dimensional time grid having a plurality of time-indexed points;

calculating a state cost metric for a state in a plurality of states of a statistical model, for a given time-indexed point, wherein the state cost metric calculation provides a most-likely predecessor state and a corresponding most-likely predecessor time-indexed point;

determining a sequence of states using the state cost metric calculation; and

determining a corresponding cumulative probability measure for the statistical model;

wherein the set of signal observations includes feature vectors, and

wherein the feature vectors are selected from at least one of: Mel-Frequency Cepstral Coefficients feature vectors, Linear Predictive Coding Coefficients, spectral density, spectral energy, noise ratios, length of sounds, relative power, and filter matches.

5. A method, comprising:

receiving a set of signal observations for a signal repetition, wherein the set of signal observations is associated with a dimension of a K-dimensional time grid having a plurality of time-indexed points, wherein the plurality of time-indexed points are determined with respect to a predetermined distance from a diagonal line through the K-dimensional time grid;

calculating a state cost metric for a state in a plurality of states of a statistical model, for a given time-indexed point, wherein the state cost metric calculation provides a most-likely predecessor state and a corresponding most-likely predecessor time-indexed point;

determining a sequence of states using the state cost metric calculation; and

determining a corresponding cumulative probability measure for the statistical model.

6. The method of claim 5 , wherein the set of observations has a first time duration, and the predetermined distance is based on a difference of the first time duration and a second time duration of another set of signal observations.

7. The method of claim 1 , wherein the determining the sequence of states comprises backtracking through the plurality of time-indexed points based on the state cost metric.

8. A method, comprising:

receiving a set of signal observations for a signal repetition, wherein the set of signal observations is associated with a dimension of a K-dimensional time grid having a plurality of time-indexed points;

calculating a state cost metric for a state in a plurality of states of a statistical model, for a given time-indexed point, wherein the state cost metric calculation provides a most-likely predecessor state and a corresponding most-likely predecessor time-indexed point, wherein the calculating is based on:

state cost metrics for states associated with candidate predecessor time-indexed points,

a probability of transitioning from each state of each candidate predecessor time-indexed point to the given state,

a respective probability of transitioning from the respective candidate predecessor time-indexed point, and

a joint probability of the observations being emitted from the state in the plurality of states;

determining a sequence of states using the state cost metric calculation; and

determining a corresponding cumulative probability measure for the statistical model.

9. The method of claim 1 , wherein the determined sequence of states determines an alignment of the set of signal observations.

10. An article of manufacture including a non-transitory computer-readable medium having instructions stored thereon that, when executed by a computing device, cause the computing device to perform operations comprising:

retrieving from memory a set of signal observations for a signal repetition, wherein the set of signal observations is associated with a dimension of a K-dimensional time grid having a plurality of time-indexed points, wherein the plurality of time-indexed points are determined with respect to a predetermined distance from a diagonal line through the K-dimensional time grid;

retrieving from memory a set of parameters for a plurality of statistical models;

calculating a state cost metric for a state in a plurality of states of a statistical model at a time-indexed point, wherein the state cost metric calculation provides a most-likely predecessor state and a corresponding most-likely predecessor time-indexed point;

determining a cumulative probability measure for the statistical model; and

determining a most likely statistical model from the plurality of statistical models.

11. An apparatus, comprising:

a processor configured to execute software applications stored in memory, the software instructions including:

calculating a state cost metric for a state of a statistical model at a time-indexed point, wherein the state cost metric calculation provides a most-likely predecessor state and a corresponding most-likely predecessor time-indexed point;

determining a sequence of states using the state cost metric calculation;

determining a corresponding cumulative probability measure for the statistical model;

calculating a cost metric associated with a possible prior state at a possible predecessor time-indexed point;

selecting a lowest cost metric for the state to determine the state cost metric;

observing dimensions that are incremented when moving from a given predecessor time-index point to a given one of the plurality of time-indexed points; and

basing the state cost metric on at least the observing.

12. The apparatus of claim 11 , further comprising a memory configured to:

store a digital representation of a set of time-sequential signal observations for a signal repetition, wherein the set of time-sequential signal observations is associated with a dimension of a k-dimensional time grid having a plurality of time-indexed points; and

store a set of parameters for a plurality of statistical models.

13. The apparatus of claim 11 , further comprising an audio receiver configured to:

receive a set of audio signal observations for a signal repetition, wherein the set of audio signal observations is associated with a dimension of a k-dimensional time grid having a plurality of time-indexed points; and

convert the set of audio signal observations into a set of analog electrical signal observations for a signal repetition.

14. The apparatus of claim 11 , further comprising an analog-to-digital converter configured to convert a set of analog electrical signal observations into a set of digital electrical signal observations.

15. The apparatus of claim 11 , wherein the processor is further configured to compute a set of features vectors based on a digital representation of a set of signal observations for a signal repetition.

16. The apparatus of claim 11 , wherein the processor is further configured to transform the sequence of states using the state cost metric calculation into a digital representation of the sequence of states.

17. The apparatus of claim 16 , further comprising a digital-to-analog converter configured to transform the digital representation of the sequence of states into an output analog electrical signal.

18. The apparatus of claim 17 , further comprising a system output interface configured to transform the output analog electrical signal into an output audio signal.