Method, apparatus and program products for retrieving moving image
Abstract
An apparatus for retrieving moving image according to the present invention has an extracting unit for extracting the information of the feature quantity of the moving image from the fixed moving image data, an extracting unit for extracting the information of the unique identification of the moving image to identify the moving image according to the information of the feature quantity of the moving image, and a retrieving unit for retrieving and identifying the fixed moving image from the object moving image for retrieving by comparing the information of the unique identification of the moving image with the information of the unique identification of the moving image beforehand extracted from the object moving image. Therefore, the information of the unique identification of the moving image for retrieving the moving image depends on the content of the moving image can be extracted, consequently, the identical moving image can be retrieved and identified from the moving image by using the information of the unique identification of the moving image at high speed.
Claims
exact text as granted — not AI-modified1 . An apparatus for retrieving predetermined moving image data from object moving image data, comprising:
first extracting means for extracting physical information of a feature quantity of each frame from the predetermined moving image data; second extracting means for extracting information of unique identification of the predetermined moving image data, based on the extracted physical information; and retrieving means for retrieving identical information to the information of unique identification of the predetermined moving image data, from an information of unique identification of the object moving image data, by comparing the information of unique identification of the predetermined moving image data with the information of unique identification of the object moving image data, wherein the information of unique identification of the predetermined moving image data and the information of unique identification of the object moving image data is time-series information created by arranging the physical information of the feature quantity extracted from each frame of the predetermined moving image data and the object moving image data, respectively, on a time axis.
2 . The apparatus for retrieving the moving image of claim 1 , wherein the retrieving means retrieves the identical information by repeatedly comparing the information of unique identification of the predetermined moving image data with all of the information of unique identification extracted from the object moving image data.
3 . The apparatus for retrieving the moving image of claim 1 , further comprising
detecting means for detecting cut points of the predetermined moving image data as information of cut points, wherein the second extracting means extracts information of unique identification in each shot unit of the predetermined moving image data which is divided by the information of the cut points; and the retrieving means retrieves the identical information to the information of unique identification of the predetermined moving image data, from the information of unique identification of the object moving image data, by comparing the information of unique identification in each shot unit extracted from the predetermined moving image data with the information of unique identification of the object moving image data which is the time-series information created by arranging the physical information of the feature quantity extracted from each frame, on the time axis.
4 . The apparatus for retrieving the moving image of claim 3 , wherein
the retrieving means compares the information of unique identification extracted from each shot unit of the predetermined moving image data, with the information of unique identification extracted from each shot unit of the object moving image data which is divided by the cut points detected in the object moving image data.
5 . The apparatus for retrieving the moving image of claim 1 , wherein the second extracting means extracts the information of unique identification by making a graph of the time-series information.
6 . The apparatus for retrieving the moving image of claim 5 , wherein the second extracting means extracts the information of unique identification by expanding or reducing the graph.
7 . The apparatus for retrieving the moving image of claim 1 , wherein
the first extracting means extracts the information of the feature quantity as a block unit of M×N pixels, where M and N are natural numbers, from the predetermined moving image data.
8 . The apparatus for retrieving the moving image of claim 1 , wherein
the second extracting means extracts a value as the information of unique identification of the predetermined moving image data, by picking out all of the time-series information, which is arranging the physical information of the feature quantity extracted from each frame, or by picking out some time-series information by uniform or irregular intervals, on the time axis, from the predetermined moving image data.
9 . The apparatus for retrieving the moving image of claim 8 , wherein
the retrieving means compares the information of unique identification of the predetermined moving image data with that of the object moving image data created by picking out all of the time-series information, which is arranging the physical information of the feature quantity of the each frame, or by picking out some time-series information by uniform or irregular intervals, on the time axis, from the object moving image data.
10 . The apparatus for retrieving the moving image of claim 1 , wherein
the physical information of the feature quantity is information of luminance and color difference (Y, Cb, Cr) or information of color (R, G, B) of each frame.
11 . The apparatus for retrieving the moving image of claim 1 , wherein the physical information of the feature quantity is one or more transformed values obtained from information of luminance and color difference (Y, Cb, Cr) or information of color (R, G, B) of each frame.
12 . The apparatus for retrieving the moving image of claim 1 , wherein
the transformed value obtained from the information of luminance and color difference (Y, Cb, Cr) of each frame is one or more of a mean value, a total value, and a difference value of the information of luminance and color difference (Y, Cb, Cr) of each frame; and the transformed value obtained from the information of color (R, G, B) of each frame is one or more of a mean value, a total value, and a difference value of the information of color (R, G, B) of each frame.
13 . The apparatus for retrieving the moving image of claim 11 , wherein
the transformed value obtained from the information of luminance and color difference (Y, Cb, Cr) of each frame is one or more of a value, a mean value, a total value, and a difference value of the weighted information of luminance and color difference (Y, Cb, Cr) of each frame; and the transformed value obtained from the information of color (R, G, B) of each frame is one or more of a value, a mean value, a total value, and a difference value of the weighted information of color (R, G, B) of each frame.
14 . The apparatus for retrieving the moving image of claim 11 , wherein the transformed value obtained from the information of luminance and color difference (Y, Cb, Cr) of each frame is a quantified value of the information of luminance and color difference (Y, Cb, Cr) of each frame; and
the transformed value obtained from the information of color (R, G, B) of each frame is a quantified value of the information of color (R, G, B) of each frame.
15 . The apparatus for retrieving the moving image of claims 12 or 13 , wherein the transformed value obtained from the information of luminance and color difference (Y, Cb, Cr) of each frame is a quantified value of one or more of the values; and
the transformed value obtained from the information of the information of color (R, G, B) of each frame is a quantified value of one or more of the values.
16 . The apparatus for retrieving the moving image of claim 1 , wherein
the information of the feature quantity is information of a feature quantity of a compressed moving image data obtained by executing a Discrete Cosine Transform operation in uncompressed moving image data, or the information of the feature quantity of the compressed moving image data extracted from compressed moving image data.
17 . The apparatus for retrieving the moving image of claim 16 , wherein
the information of the feature quantity of the compressed moving image, as the physical information of the future quantity of each frame, is Discrete Cosine Transform coefficients of each frame or information of motion compensation of each frame.
18 . The apparatus for retrieving the moving image of claim 16 , wherein
the information of the feature quantity of the compressed moving image, as the physical information of the feature quantity of each frame, is one of a value, a mean value, a total value, and a difference value of the Discrete Cosine Transform coefficients of each frame, and a vector or a value of the vector which is at least one of a motion vector, a mean motion vector between frames before and after, a total motion vector, a difference motion vector, and a motion vector on a whole frame obtained from the information of the motion compensation.
19 . The apparatus for retrieving the moving image of claim 16 , wherein
the information of the feature quantity of the compressed, moving image, as the physical information of the feature quantity of each frame, is one of a value, a mean value, a total value, and a difference value of weighted Discrete Cosine Transform coefficients of each frame, and a vector or a value of the vector which is at least one of a motion vector, a mean motion vector between frames before and after, a total motion vector, a difference motion vector, and a motion vector on a whole frame obtained from weighted information of the motion compensation.
20 . The apparatus for retrieving the moving image of claim 16 , wherein
the information of the feature quantity of the compressed moving image is only a direct current element of the Discrete Cosine Transform coefficients of each frame, as the physical information of the future quantity of each frame.
21 . The apparatus for retrieving the moving image of claim 16 , wherein
the information of the feature quantity of the compressed moving image, as the physical information of the feature quantity of each frame, is a quantified value of one of the Discrete Cosine Transform coefficients of each frame, a value obtained from the Discrete Cosine Transform coefficients of each frame, a value of the motion compensation, and a value obtained from the value of the motion compensation.
22 . The apparatus for retrieving the moving image of claim 1 , wherein
the information of the feature quantity is the combination of: information of luminance and color difference (Y, Cb, Cr) or the information of color (R, G, B) of each frame, and information of the feature quantity of the compressed moving image which is obtained from an uncompressed moving image data by executing a Discrete Cosine Transform operation or which is obtained by extracting from a compressed moving image data.
23 . An apparatus for monitoring television commercials to retrieve a predetermined moving image data such as a television commercial from an object moving image data such as the television broadcast, wherein the apparatus for monitoring television commercials is the apparatus for retrieving the moving image of claim 1 .
24 . The apparatus for monitoring television commercials of claim 23 , further comprising:
storing means for storing a result of retrieving by the retrieving means; and output means for displaying or printing the result of retrieving stored by the storing means.
25 . A method for retrieving predetermined moving image data from object moving image data, comprising:
a first extracting step for extracting physical information of a feature quantity of each frame from the predetermined moving image data; a second extracting step for extracting information of unique identification of the predetermined moving image data, based on the extracted physical information; and a retrieving step for retrieving an identical information to the information of unique identification of the predetermined moving image data, from an information of unique identification of the object moving image data, by comparing the information of unique identification of the predetermined moving image data with the information of unique identification of the object moving image data, wherein the information of unique identification of the predetermined moving image data and the information of unique identification of the object moving image data is time-series information created by arranging the physical information of the feature quantity extracted from each frame of the predetermined moving image data and the object moving image data, respectively, on a time axis.
26 . The method for retrieving the moving image of claim 25 , wherein
the retrieving step retrieves the identical information by repeatedly comparing the unique identification of the predetermined moving image data with all of the information of unique identification extracted from the object moving image data.
27 . The method for retrieving the moving image of claim 25 , further comprising:
a detecting step for detecting cut points of the predetermined moving image data as information of cut points, wherein the second extracting step extracts information of unique identification in each shot unit of the predetermined moving image data which is divided by the information of the cut points; and the retrieving step retrieves the identical information to the information of unique identification of the predetermined moving image data, from the information of unique identification of the object moving image data. by comparing the information of unique identification in each shot unit extracted from the predetermined moving image data with the information of unique identification of the object moving image data which is the time-series information created by arranging the physical information of the feature quantity extracted from each frame, on the time axis.
28 . The method for retrieving the moving image of claim 27 , wherein
the retrieving step compares the information of unique identification extracted from each shot unit of the predetermined moving image data with the information of unique identification extracted from each shot unit of the object moving image data which is divided by the cut points detected in the object moving image data.
29 . The method for retrieving the moving image of claim 25 , wherein the second extracting step extracts the information of unique identification by making a graph of the time-series information.
30 . The method for retrieving the moving image of claim 29 , wherein the second extracting step extracts the information of unique identification by expanding or reducing the graph.
31 . The method for retrieving the moving image of claim 25 , wherein
the first extracting step extracts the information of the feature quantity as a block unit of M×N pixels, where M and N are natural numbers, from the predetermined moving image data.
32 . A method for monitoring television commercials to retrieve a predetermined moving image data such as a television commercial from an object moving image data such as the television broadcast, wherein the method for monitoring television commercials is the method for retrieving the moving image of claim 25 .
33 . The method for monitoring television commercials of claim 32 , further comprising:
storing a result of retrieving by the retrieving step; and displaying or printing the result of retrieving stored by the storing step.
34 . A computer program product comprising a computer readable medium having computer readable program code embodied thereon, the program code being adapted to carry out all steps of claim 25 .Join the waitlist — get patent alerts
Track US2009049103A1 — get alerts on status changes and closely related new filings.
We store only your email — no account needed. See our privacy policy.