Claims Analysis
Total Claims: 9
Claim 1
Independent Claim
200 words
A device for recognizing sequences in a signal to be analyzed wherein said signal to be analyzed is produced continuously and is constituted by multi-dimensional vectors, said device comprising, for each of said sequences to be recognized; means for providing sequence vectors representing said sequence to be recognized; means for providing a masking vector determined from said sequence vectors; first means for correlating of said multi-dimensional vectors constituting said signal to be analyzed with said produced masking vector wherein the output of said first means for correlation provides a series of values corresponding to the degree of similarity of said multi-dimensional vectors representing said signal with said masking vector; reference sequence producing means for providing a reference sequence determined from said vectors forming said sequence to be recognized; second correlation means for correlating a sequence of said series of values with said reference sequence whereby said second correlation means produces values corresponding to the degree of similarity of said series of values and said reference sequence; and means for deciding on the validity of the recognition, by comparison of said values corresponding to the degree of similarity of said series of values and said reference sequence with a threshold value.
Claim 2
Independent Claim
200 words
A device for recognizing sequences in a signal to be analyzed wherein said signal to be analyzed is produced continuously and is constituted by multi-dimensional vectors, said device comprising, for each of said sequences to be recognized; means for providing sequence vectors representing said sequence to be recognized; means for providing a masking vector determined from said sequence vectors; first means for correlating of said multi-dimensional vectors constituting said signal to be analyzed with said produced masking vector wherein the output of said first means for correlation provides a series of values corresponding to the degree of similarity of said multi-dimensional vectors representing said signal with said masking vector; reference sequence producing means for providing a reference sequence determined from said vectors forming said sequence to be recognized; second correlation means for correlating a sequence of said series of values with said reference sequence whereby said second correlation means produces values corresponding to the degree of similarity of said series of values and said reference sequence; and means for deciding on the validity of the recognition, by comparison of said values corresponding to the degree of similarity of said series of values and said reference sequence with a threshold value.
Claim 3
Dependent Claim
26 words
A device according to claim 1, wherein said vectors representing said signal are provided by an information compression means, from said multi-dimensional vectors of said signal.
Claim 4
Dependent Claim
17 words
A device according to claim 2, wherein said information compression means includes means for providing average values.
Claim 5
Dependent Claim
25 words
A device according to claim 1, wherein the first correlation means and/or the second correlation means include means for the computation of an Euclidean norm.
Claim 6
Dependent Claim
20 words
A device according to claim 1, including means for the adapting of said threshold value to the permissible noise level.
Claim 7
Dependent Claim
26 words
A device according to claim 1, including means for the transformation of said degree of similarity, for example as a function of the maximum threshold value.
Claim 8
Dependent Claim
29 words
A device according to claim 1, including means for the determining of said masking vector and/or said reference sequence, using a learning rule that minimizes the resistance to noise.
Claim 9
Dependent Claim
62 words
A device according to claim 7, wherein said means for the determining of said masking vector include means to obtain the average of said vectors representing said sequence to be recognized, and wherein said means for the determining of said reference sequence include means for the correlation of each of said vectors representing said sequence to be recognized with said masking vector.