Method for the scaling of the indexing data of a multimedia document
First Claim
Patent Images
1. A method for the scaling of indexing data D=(dn, n=1 to N) of a multimedia document wherein mainly the method comprises the following steps which consist:
- a) at the time t, in grouping the data D in distinct and consecutive groups Dj respectively sized n′
j, j varying from 1 to J and respectively scaling each group Dj by a value d′
j according to at least one determined scaling method C and in storing the data D′
=(d′
j,j=1 to J) thus obtained, b) subsequently, at the time t′
>
t, when the number of data resulting from the previous scaling operation is too great, in grouping the data D′
in distinct and consecutive groups D′
k respectively sized Jk, k varying from 1 to K and respectively scaling each group D′
k by a value d″
k according to a rescaling method C′
compatible with the scaling method C in such a way that each of the data d″
k is equivalent to that obtained by applying the scaling method C directly to distinct and consecutive groups of data D respectively sized n″
k, n″
k being the sum of the n′
j values of the group D′
k and in storing the scaled data D″
=(d″
k, k=1 to K).
6 Assignments
0 Petitions
Accused Products
Abstract
A method for scaling of indexing data D=(dn, n=1 to N) of a multimedia document having the following steps: (a) at the time t, in grouping the data D in distinct and consecutive groups Dj respectively sized n′j, j varying from 1 to J and in respectively scaling each group Dj by a value d′j according to at least one determined scaling method C and in storing the data D′=(d′j,j=1 to J) thus obtained, (b) subsequently, at the time t′>t, when the number of data resulting from the previous scaling is too great, in grouping the data D′ in distinct and consecutive groups D′k respectively sized Jk, k varying from 1 to K and respectively scaling each group D′k by a value d″k according to a rescaling method C′ compatible with the scaling method C in such a way that each of the data d″k is equivalent to that obtained by applying the scaling method C directly to distinct and consecutive groups of data D respectively sized n″k, n″k being the sum of the n′j values of the group D′k and in storing the scaled data D″=(d″k, k=1 to K).
-
Citations
20 Claims
-
1. A method for the scaling of indexing data D=(dn, n=1 to N) of a multimedia document wherein mainly the method comprises the following steps which consist:
-
a) at the time t, in grouping the data D in distinct and consecutive groups Dj respectively sized n′
j, j varying from 1 to J and respectively scaling each group Dj by a value d′
j according to at least one determined scaling method C and in storing the data D′
=(d′
j,j=1 to J) thus obtained,b) subsequently, at the time t′
>
t, when the number of data resulting from the previous scaling operation is too great, in grouping the data D′
in distinct and consecutive groups D′
k respectively sized Jk, k varying from 1 to K and respectively scaling each group D′
k by a value d″
k according to a rescaling method C′
compatible with the scaling method C in such a way that each of the data d″
k is equivalent to that obtained by applying the scaling method C directly to distinct and consecutive groups of data D respectively sized n″
k, n″
k being the sum of the n′
j values of the group D′
k and in storing the scaled data D″
=(d″
k, k=1 to K).- View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20)
each datum dn is weighted by a weight wn, each datum d′
j is weighted by a weight w′
j, each of these weights being equal to the sum of the weights of the corresponding data of the groups Dj,each datum d″
k is weighted by a weight w″
k, each of these weights being equal to the sum of the weights of the corresponding data of the groups D′
k,and wherein the weight of each data is associated with said data.
-
-
5. A method according to claim 1 wherein, prior to the storage of the scaled data, a header comprising at least one label specifying the scaling method is associated with the scaled data.
-
6. A method according to claim 5, wherein the header advantageously comprises the number of data before the scaling and/or the number of data obtained after the scaling.
-
7. A method according to claim 1, wherein the determined scaling method C is the method C7 based on the histogram of each data group Dj according to predefined categories and the rescaling method C″
- is the method C7′
based on the computation of the sum, term by term, of groups of histograms of D′
.
- is the method C7′
-
8. A method according to claim 1, wherein the data D are series of scalar values or vectors.
-
9. A method according to claim 1, wherein the determined scaling method C is the method C4 which consists of the random choice of a datum from each group of data Dj and the determined rescaling method C′
- is then the method C4′
which consists of the random choice of a datum from each group of data D′
k.
- is then the method C4′
-
10. A method according to claim 1, wherein the data d″
-
k are equal to those obtained by applying the scaling method C directly to distinct and consecutive groups of data D respectively sized n″
k.
-
k are equal to those obtained by applying the scaling method C directly to distinct and consecutive groups of data D respectively sized n″
-
11. A method according to claim 10, wherein the determined scaling method C is the method C3 based on the computation of the mean of each group of data Dj and the rescaling method C′
- is the method C3′
based on the computation of the mean of each group of data D′
k.
- is the method C3′
-
12. A method according to the claim 10, wherein the determined scaling method C is the method C1 based on the computation of the minimum of each group of data Dj and the rescaling method C′
- is the method C1′
based on the computation of the minimum of each group of data D′
k.
- is the method C1′
-
13. A method according to the claim 10, wherein the determined scaling method C is the method C2 based on the computation of the maximum of each group of data Dj and the rescaling method C′
- is the method C2′
based on the computation of the maximum of each group of data D′
k.
- is the method C2′
-
14. A method according to the claim 10, wherein the determined scaling method C is the method C5 based on the choice of the first datum from each group of data Dj and the rescaling method C′
- is the method C5′
based on the choice of the first datum from each group of data D′
k.
- is the method C5′
-
15. A method according to the claim 10, wherein the determined scaling method C is the method C6 based on the computation of the last datum from each group of data Dj and the rescaling method C′
- is the method C6′
based on the choice of the last datum from each group of data D′
k.
- is the method C6′
-
16. A method according to the claim 10, wherein the determined scaling method C is the method C3 based on the computation of the mean of each group of data Dj, then the method C8 based on the computation of the variance of each group of data Dj and the rescaling method C′
- is the method C8′
based on the means of the groups of data D′
k resulting from the scaling according to C8 and the variances of the groups of data D′
k resulting from the scaling according to C3.
- is the method C8′
-
17. A method according to the claim 10, wherein sizes n′
- j and Jk are all powers of two.
-
18. A method according to claim 17, wherein the determined scaling method C is the method C3 based on the computation of the mean of each group of data Dj and then the method C9 based on a decomposition of the variance of each group of data Dj into a series of coefficients each describing the variability at a particular scale, and the rescaling method C′
- is then the method C9′
, based on the means of the groups of data D′
k resulting from the scaling according to C9, and a decomposition of the variance of the groups of data D′
k resulting from the scaling according to C3.
- is then the method C9′
-
19. A method according to the claim 10, the data being vectors, wherein the determined scaling method C is the method C3 based on the computation of the mean of each group of data Dj and then the method C10 based on the computation of the covariance of each group of data Dj and the rescaling method C′
- is the method C10′
based on the computation of the means of the groups of data D′
k resulting from the scaling according to C10 and of the covariances of the groups of data D′
k resulting from the scaling according to C3.
- is the method C10′
-
20. A method according to the claim 10, the data being vectors, wherein the determined scaling method C is the method C3 based on the computation of the mean of each group of data Dj, and then the method C11 based on the computation of the sum of the terms of the diagonal of the covariance matrix of each group of data Dj and the rescaling method C′
- is the method C11′
based on the computation of the mean of each group of data D′
k resulting form the scaling according to C11 and the sum of the terms of the diagonal of the covariance matrix each group of data D′
k resulting from the scaling according to C3.
- is the method C11′
Specification
-
- Resources
Litigation Campaign Assessment
Thank you for your request. You will receive a custom alert email when the Litigation Campaign Assessment is available.
×
-
Current AssigneeMusicqubed Innovations LLC (Ascend Innovation Management, LLC), Centre National De La Recherche Scientifique
-
Original AssigneeCentre National De La Recherche Scientifique, Orange S.A.
-
InventorsDe Cheveigne, Alain
-
Primary Examiner(s)Mizrahi, Diane D.
-
Assistant Examiner(s)Punit, Prakash C.
-
Application NumberUS09/964,250Publication NumberTime in Patent Office818 DaysField of Search707/1-3, 707/4, 707/100, 707/102, 707/104.1, 382/173US Class Current1/1CPC Class CodesG06F 16/40 of multimedia data, e.g. sl...G06F 16/45 Clustering; ClassificationY10S 707/99933 Query processing, i.e. sear...Y10S 707/99943 Generating database or data...Y10S 707/99945 Object-oriented database st...
