Signal processing method and image-voice processing apparatus for measuring similarities between signals

• US 6,710,822 B1
• Filed: 10/12/2000
• Issued: 03/23/2004
• Est. Priority Date: 02/15/1999
• Status: Expired due to Fees

• Current Assignee: Sony Corporation (Sony Group Corp.)
• Original Assignee: Sony Corporation (Sony Group Corp.)
• Inventors: Walker, Toby, Matsubara, Hiroshi
• Primary Examiner(s): Kostak, Victor R.
• Application Number:
  US09/673,232
• Time in Patent Office: 1,258 Days
• Field of Search: 348/722, 348/700, 348/180, 345/723-726, 382/170, 382/224, 382/232
• US Class Current: 348/722
• CPC Class Codes: G06F 16/683 using metadata automatically derived from the content
• CPC Class Codes: G06F 16/7834 using audio features
• CPC Class Codes: G06F 16/785 using colour or luminescence
• CPC Class Codes: G06V 20/41 Higher-level, semantic clustering, classification or understanding of video scenes, e.g. detection, labelling or Markovian modelling of sport events or news items segmenting video sequences G06V20/49
• CPC Class Codes: G11B 27/105 of operating discs
• CPC Class Codes: G11B 27/28 by using information signals recorded by the same method as the main recording G11B27/22 takes precedence

Abstract

In step S1 the image-voice processing apparatus 10 divides segments into sub-segments that will be candidates for r segments, and in step S2 groups together similar sub-segments out of the sub-segments obtained. And in step S3 the image-voice processing apparatus 10 selects object groups for signatures, and then after selecting in step S4 a r segment from each of the groups selected, computes in step S5 the weight for each of the r segments obtained.

Claims

1. A signal processing method that extracts signatures defined by the representative segments which are sub-segments that represent the contents of segments constituting signals supplied out of the sub-segments contained in said segments and a weighting function that allocates weights to these representative segments, said signal processing method comprising:

a group selection step that selects object groups for said signatures out of the groups obtained by a classification based on an arbitrary attribute of said sub-segment;

a representative segment selection step that selects a representative segment out of the groups selected in said group selection step; and

a weight computing step that computes the weight for the representative segment obtained in said selection step,

wherein the results of comparison obtained by comparing the similarity of each representative segment of mutually different segments and the weight given to each representative segment computed in said computing step are used to compare the similarity of mutually different segments.

2. A signal processing method according to claim 1 further comprising:

a segment fragmentation step that fragments said segment into a plurality of sub-segments that will be candidates for said representative segments; and

a grouping step that classifies and groups said sub-segments based on the mutual similarity for said attribute from among the sub-segments obtained in said segment fragmentation step; and

wherein said group selection step selects arbitrary object groups for said signatures from out of the groups obtained through said segment fragmentation step and said grouping step.

3. A signal processing method according to claim 1 wherein said signals are at least one of image signals and voice signals contained in video data.

4. A signal processing method according to claim 1 wherein said weight computing step computes said weight based on the total number of sub-segments contained in the groups corresponding to the representative segments.

5. A signal processing method according to claim 1 wherein said segments constitute an arbitrary stream of said signals.

6. A signal processing method according to claim 1 wherein said sub-segments constitute an arbitrary stream contained in said segments.

7. A signal processing method according to claim 1 wherein said segments constitute a scene consisting of segments formed by a stream of successive frames or chronologically successive segments forming part of said signal.

8. A signal processing method according to claim 1 wherein said segments constitute a program representing the whole of the frames constituting said signals or said signals.

9. An image-voice processing apparatus that extracts signatures defined by the representative segments which are image and/or voice sub-segments that represent the contents of image and/or voice segments constituting video signals supplied out of the image and/or voice sub-segments contained in said image and/or voice segments and a weighting function that allocates weight to these representative segments, said image-voice processing apparatus comprising:

means that selects object groups for said signatures out of the groups obtained by a classification based on an arbitrary attribute of said image and/or voice sub-segments,

means that selects a representative segment out of these selected groups; and

means that computes a weight for the representative segment obtained thereby,

wherein the results of comparison obtained by comparing the similarity among the respective representative segments of mutually different image and/or voice segments and the weight given to each representative segment are used to compare the similarity among mutually different image and/or voice segments.

10. An image-voice processing apparatus according to claim 9 wherein said image and/or voice segments are divided into a plurality of image and/or voice sub-segments which will be candidates for said representative segments, said image and/or voice sub-segments are classified and grouped together based on their mutual similarity as said attribute from among the image and/or voice sub-segments obtained thereby, and arbitrary groups to be objects of said signatures are selected.

11. An image-voice processing apparatus according to claim 9 wherein said weight is computed based on the total number of image and/or voices sub-segments contained in groups corresponding to the representative segments.

12. An image-voice processing apparatus according to claim 9 wherein said image and/or voice segments are an arbitrary stream of said video signals.

13. An image-voice processing apparatus according to claim 9 wherein said image and/or voice sub-segments are an arbitrary stream contained in said image and/or voice segments.

14. An image-voice processing apparatus according to claim 9 wherein said image and/or voice sub-segments constitute a scene consisting of a stream of successive image and/or voice frames constituting said video signals or chronologically successive image and/or voice segments.

15. An image-voice processing apparatus according to claim 9 wherein said image and/or voice sub-segments are image and/or voice frames constituting said video signals or a program representing the whole video signals.