초록 ▼

In one form of the invention, a computer database storage system is disclosed, comprising a data storage medium adapted to store a plurality of pieces of information, at least one piece of data stored in the data storage medium, and at least one rule stored in the data storage medium, each said at l

In one form of the invention, a computer database storage system is disclosed, comprising a data storage medium adapted to store a plurality of pieces of information, at least one piece of data stored in the data storage medium, and at least one rule stored in the data storage medium, each said at least one rule comprising a premise, an action, wherein the action is performed if the premise is determined to be true, an alternate action, wherein the alternate action is performed if the premise is determined to be false, and a trigger, wherein the trigger causes evaluation of the premise upon the occurrence of a predetermined event.

대표청구항 ▼

In one form of the invention, a computer database storage system is disclosed, comprising a data storage medium adapted to store a plurality of pieces of information, at least one piece of data stored in the data storage medium, and at least one rule stored in the data storage medium, each said at l

In one form of the invention, a computer database storage system is disclosed, comprising a data storage medium adapted to store a plurality of pieces of information, at least one piece of data stored in the data storage medium, and at least one rule stored in the data storage medium, each said at least one rule comprising a premise, an action, wherein the action is performed if the premise is determined to be true, an alternate action, wherein the alternate action is performed if the premise is determined to be false, and a trigger, wherein the trigger causes evaluation of the premise upon the occurrence of a predetermined event. irs of frames to determine a corresponding matching score for each sub-sequence indicative of the similarity between the aligned pairs of frames within each sub-sequence; first comparing means for comparing the matching scores for the sub-sequences to identify a worst matching portion of the sequence of aligned pairs of frames; second comparing means for comparing said average frame score and a matching score of said worst matching portion with stored model data which defines consistent training speech signals; and means for determining if said first and second input speech signals are consistent with each other from a comparison result output by said second comparing means. 7. A consistency checking method, comprising the steps of: receiving a first sequence of frames representative of a first speech signal and a second sequence of frames representative of a second signal; matching the frames representative of the first speech signal with frames representative of the second speech signal to determine a sequence of aligned pairs of frames, with each aligned pair of frames including a frame from said first sequence and a frame from said second sequence, and to determine a matching score associated with the sequence of aligned pairs of frames indicative of the similarity between the first and second sequences of frames; calculating an average frame score by dividing said matching score by the number of aligned pairs of frames in said sequence of aligned pairs of frames; processing sub-sequences of said sequence of aligned pairs of frames to determine a corresponding matching score for each sub-sequence indicative of the similarity between the aligned pairs of frames within each sub-sequence; a first comparing step of comparing the matching scores for the sub-sequences to identify a worst matching portion of the sequence of aligned pairs of frames; a second comparing step of comparing said average frame score and a matching score of said worst matching portion with stored model data which defines consistent training speech signals; and determining if said first and second input speech signals are consistent with each other from a comparison result output by said second comparing step. 8. A method according to claim 7, wherein said processing step is arranged to process sub-sequences of equal size. 9. A method according to claim 7, wherein said model comprises the average of the average frame scores and the average of the matching scores of the worst matching portions identified in a set of training speech signals which are known to be consistent. 10. A method according to claim 7, wherein said model models the variation in the average frame scores and the matching scores of the worst matching portions and/or the correlation between the average frame scores and the matching scores of the worst matching portions in a set of training speech signals which are known to be consistent. 11. A method according to claim 7, wherein said matching step uses a dynamic programming technique to generate said matching score. 12. A consistency checking method of checking the consistency between a first sequence of frames representative of a first speech signal and a second sequence of frames representative of a second speech signal using a matching score and a sequence of aligned pairs of frames generated by a matching process performed on said first and second sequences of frames, the method comprising the steps of: calculating an average frame score by dividing said matching score by the number of aligned pairs of frames in said sequence of aligned pairs of frames; processing sub-sequences of said sequence of aligned pairs of frames to determine a corresponding matching score for each sub-sequence indicative of the similarity between the aligned pairs of frames within each sub-sequence; a first comparing step of comparing the matching scores for the sub-sequences to identify a worst matching portion of the sequence of aligne