Sponsorship credit period identification apparatus, sponsorship credit period identification method and program
View Patent ↗A credit segment identifying device includes an extracting unit. The extracting units extracts, from a first speech signal, a plurality of first partial speech signals. Each of the plurality of first partial speech signals is a part of the first speech signals and shifted from each other in time direction. An identifying unit identifies a credit segment in the first speech signal by determining whether each of the first partial speech signals includes a credit according to an association between each of second partial signals and the presence/absence of a credit. The each of second partial signals is extracted from a second speech signal.
1 . A credit segment identifying device comprising:
an extractor configured to extract a plurality of first partial speech signals from a first speech signal, the first partial speech signals each being a part of the first speech signal and shifted from each other in time direction;
an identifier configured to identify a credit segment in the first speech signal by determining whether each of the first partial speech signals includes a credit according to a first set of partial speech signal comprising a previously set term and identifying a second set of speech signals not being in the same category as the first set of partial speech signals extracted from a second speech signal; wherein
identifying the credit segment further comprising:
comparing the each of the first partial speech signals to a predetermined threshold; and
generating a binary time-series signal on the each of the first partial speech signals that exceeds the predetermined threshold;
wherein the extractor extracts a plurality of first still images corresponding to the first partial speech signals from a first video signal corresponding to the first speech signal, and the identifier identifies a credit segment in the first speech signal and the first video signal by determining whether each pair of the first partial speech signal and the first still image includes a credit according to each of the second partial speech signals and an association between a second still image extracted from a second video signal corresponding to the second speech signal and corresponding to each of the second partial speech signals and the presence/absence of a credit.
2 . The credit segment identifying device according to claim 1 , wherein the identifier determines whether each of the first partial speech signals includes a credit using an identifier model which has learned each of the second partial speech signals and the presence/absence of a credit.
3 . The credit segment identifying device according to claim 2 , wherein the extractor extracts a plurality of first still images corresponding to the first partial speech signals from a first video signal corresponding to the first speech signal, and the identifier identifies a credit segment in the first speech signal and the first video signal by determining whether each pair of the first partial speech signal and the first still image includes a credit according to each of the second partial speech signals and an association between a second still image extracted from a second video signal corresponding to the second speech signal and corresponding to each of the second partial speech signals and the presence/absence of a credit.
4 . The credit segment identifying device according to claim 1 , wherein whether the second partial speech signal includes the term is determined according to speech recognition carried out to the second partial speech signal as a target.
5 . The credit segment identifying device according to claim 4 , wherein the identifier determines whether each of the first partial speech signals includes a credit using an identifier model which has learned each of the second partial speech signals and the presence/absence of a credit.
6 . The credit segment identifying device according to claim 4 , wherein the extractor extracts a plurality of first still images corresponding to the first partial speech signals from a first video signal corresponding to the first speech signal, and the identifier identifies a credit segment in the first speech signal and the first video signal by determining whether each pair of the first partial speech signal and the first still image includes a credit according to each of the second partial speech signals and an association between a second still image extracted from a second video signal corresponding to the second speech signal and corresponding to each of the second partial speech signals and the presence/absence of a credit.
7 . A method for identifying a credit segment, the method comprising:
extracting, by an extractor, from a first speech signal, a plurality of first partial speech signals each being a part of the first speech signal and shifted from each other in time direction;
identifying, by an identifier, a credit segment in the first speech signal by determining whether a credit is included in each of the partial speech signals according to a first set of partial speech signal comprising a previously set term and identifying a second set of speech signals not being in the same category as the first set of partial speech signals extracted from a second speech signal; wherein
identifying the credit segment further comprising:
comparing the each of the first partial speech signals to a predetermined threshold; and
generating a binary time-series signal on the each of the first partial speech signals that exceeds the predetermined threshold;
wherein the extractor extracts a plurality of first still images corresponding to the first partial speech signals from a first video signal corresponding to the first speech signal, and the identifier identifies a credit segment in the first speech signal and the first video signal by determining whether each pair of the first partial speech signal and the first still image includes a credit according to each of the second partial speech signals and an association between a second still image extracted from a second video signal corresponding to the second speech signal and corresponding to each of the second partial speech signals and the presence/absence of a credit.
8 . The method according to claim 7 , wherein the identifier determines whether each of the first partial speech signals includes a credit using an identifier model which has learned each of the second partial speech signals and the presence/absence of a credit.
9 . The method according to claim 8 , wherein the extractor extracts a plurality of first still images corresponding to the first partial speech signals from a first video signal corresponding to the first speech signal, and the identifier identifies a credit segment in the first speech signal and the first video signal by determining whether each pair of the first partial speech signal and the first still image includes a credit according to each of the second partial speech signals and an association between a second still image extracted from a second video signal corresponding to the second speech signal and corresponding to each of the second partial speech signals and the presence/absence of a credit.
10 . The method according to claim 7 , wherein the second partial speech signal is a speech signal including a previously set term, and whether the second partial speech signal includes the term is determined according to speech recognition carried out to the second partial speech signal as a target.
11 . The method according to claim 10 , wherein the identifier determines whether each of the first partial speech signals includes a credit using an identifier model which has learned each of the second partial speech signals and the presence/absence of a credit.
12 . The method according to claim 10 , wherein the extractor extracts a plurality of first still images corresponding to the first partial speech signals from a first video signal corresponding to the first speech signal, and the identifier identifies a credit segment in the first speech signal and the first video signal by determining whether each pair of the first partial speech signal and the first still image includes a credit according to each of the second partial speech signals and an association between a second still image extracted from a second video signal corresponding to the second speech signal and corresponding to each of the second partial speech signals and the presence/absence of a credit.
13 . A computer-readable non-transitory recording medium storing a computer-executable program instructions that when executed by a processor cause a computer system to:
extract by an extractor, from a first speech signal, a plurality of first partial speech signals each being a part of the first speech signal and shifted from each other in time direction; and
identify, by an identifier, a credit segment in the first speech signal by determining whether a credit is included in each of the partial speech signals according to a first set of partial speech signal comprising a previously set term and identifying a second set of speech signals not being in the same category as the first set of partial speech signals extracted from a second speech signal; wherein
identifying the credit segment further comprising:
comparing the each of the first partial speech signals to a predetermined threshold; and
generating a binary time-series signal on the each of the first partial speech signals that exceeds the predetermined threshold;
wherein the extractor extracts a plurality of first still images corresponding to the first partial speech signals from a first video signal corresponding to the first speech signal, and the identifier identifies a credit segment in the first speech signal and the first video signal by determining whether each pair of the first partial speech signal and the first still image includes a credit according to each of the second partial speech signals and an association between a second still image extracted from a second video signal corresponding to the second speech signal and corresponding to each of the second partial speech signals and the presence/absence of a credit.
14 . The computer-readable non-transitory recording medium of claim 13 , wherein the identifier determines whether each of the first partial speech signals includes a credit using an identifier model which has learned each of the second partial speech signals and the presence/absence of a credit.
15 . The computer-readable non-transitory recording medium of claim 13 , wherein the second partial speech signal is a speech signal including a previously set term, and whether the second partial speech signal includes the term is determined according to speech recognition carried out to the second partial speech signal as a target.
16 . The computer-readable non-transitory recording medium of claim 15 , wherein the identifier determines whether each of the first partial speech signals includes a credit using an identifier model which has learned each of the second partial speech signals and the presence/absence of a credit.
17 . The computer-readable non-transitory recording medium of claim 15 , wherein the extractor extracts a plurality of first still images corresponding to the first partial speech signals from a first video signal corresponding to the first speech signal, and the identifier identifies a credit segment in the first speech signal and the first video signal by determining whether each pair of the first partial speech signal and the first still image includes a credit according to each of the second partial speech signals and an association between a second still image extracted from a second video signal corresponding to the second speech signal and corresponding to each of the second partial speech signals and the presence/absence of a credit.