Information processing apparatus, information processing method, and computer program

• US 7,777,121 B2
• Filed: 08/08/2008
• Issued: 08/17/2010
• Est. Priority Date: 08/21/2007
• Status: Expired due to Fees

• Current Assignee: Sony Corporation (Sony Group Corp.)
• Original Assignee: Sony Corporation (Sony Group Corp.)
• Inventors: Asano, Yasuharu
• Primary Examiner(s): Warren; David S.
• Application Number:
  US12/189,081
• Publication Number:
  US 20090055006A1
• Time in Patent Office: 739 Days
• Field of Search: 84600-602, 707/104.1
• US Class Current: 84/600
• CPC Class Codes: G06F 16/435 Filtering based on additional data, e.g. user or group profiles
• CPC Class Codes: G06F 16/4387 by the use of playlists
• 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

An information processing apparatus performing a process for generating a playlist defining a reproduction sequence of contents includes: a model information holding part storing therein probability models corresponding to a time series pattern of content feature volumes being feature information about contents; a content feature extracting part acquiring a content feature volume corresponding to each of contents to be reproduced; a playlist generating part comparing a time series pattern of the content feature volumes extracted in the content feature extracting part corresponding to each of permutation patterns of a reproduction sequence of contents to be reproduced with a probability model held in the model information holding part, and generating a playlist in which a reproduction sequence of contents is set in accordance with a time series pattern of content feature volumes most analogous to the probability model; and a content reproducing part reproducing contents in accordance with the generated playlist.

Claims

1. An information processing apparatus which performs a process for generating a playlist that defines a reproduction sequence of contents comprising:

a model information holding part configured to store therein a plurality of probability models each corresponding to a time series pattern of content feature volumes, the content feature volume being feature information about a content;

a content feature extracting part configured to acquire a content feature volume corresponding to each of a plurality of contents to be reproduced;

a playlist generating part configured to compare a time series pattern of the content feature volumes extracted in the content feature extracting part corresponding to each of a plurality of permutation patterns of a reproduction sequence of contents to be reproduced with a probability model held in the model information holding part, and generates a playlist in which a reproduction sequence of contents is set in accordance with a time series pattern of content feature volumes most analogous to the probability model;

a content reproducing part configured to generate and store content reproduction history information in a content reproduction history information holding part and reproduce contents in accordance with the playlist generated in the playlist generating part; and

a model adapting part configured to perform a probability model update process in which the model adapting part generates a time series of content feature volumes corresponding to a reproduction history matched with content reproduction history information stored in the content reproduction history information holding part and adapts a parameter included in the generated time series of content feature volumes corresponding to the reproduction history to a probability model held in the model information holding part, whereby the model adapting part approximates the probability model to the time series of content feature volumes corresponding to the reproduction history.

2. The information processing apparatus according to claim 1,

wherein the content reproducing part is configured to perform a process in which the content reproducing part generates and stores content reproduction history information in a content reproduction history information holding part,

the information processing apparatus further includes:

an optimal model selecting part configured to compare a time series pattern of content feature volumes corresponding to a reproduction history matched with content reproduction history information stored in the content reproduction history information holding part with a plurality of probability models held in the model information holding part, and to select a probability model most analogous to the time series pattern of content feature volumes corresponding to the reproduction history as an optimal model,

wherein the playlist generating part compares a time series pattern of the content feature volumes extracted in the content feature extracting part corresponding to each of a plurality of permutation patterns of a reproduction sequence of contents to be reproduced with the optimal model, and generates a playlist in which a reproduction sequence of contents corresponding to a time series pattern of content feature volumes most analogous to the optimal model.

3. The information processing apparatus according to claim 1, further comprising:

a model creating part configured to generate a probability model to which sample data containing a reproduction sequence of contents is adapted, to perform a learning process for the created probability model to create a plurality of different probability models, and to store the created probability model in the model information holding part.

4. The information processing apparatus according to claim 1, further comprising:

a reproduction content selecting part configured to select a plurality of contents configured of a time series pattern of content feature volumes analogous to a probability model held in the model information holding part as contents to be reproduced.

5. The information processing apparatus according to claim 1,

wherein the model information holding part is configured to store therein a probability model in accordance with a hidden Markov model (HMM) as a probability model corresponding to a time series pattern of content feature volumes, the content feature volume being feature information about a content.

6. The information processing apparatus according to claim 1,

wherein the content feature volume includes at least any one of feature information about a data signal configuring a content and metainformation corresponding to a content.

7. An information processing method which performs a process for generating a playlist that defines a reproduction sequence of contents in an information processing apparatus, the method comprising the steps of:

extracting a content feature wherein a content feature extracting part acquires a content feature volume corresponding to each of a plurality of contents to be reproduced;

generating a playlist wherein a playlist generating part compares a time series pattern of the content feature volumes extracted in the step of extracting a content feature corresponding to each of a plurality of permutation patterns of a reproduction sequence of contents to be reproduced with a probability model held in advance in a model information holding part, and generates a playlist in which a reproduction sequence of contents is set in accordance with a time series pattern of content feature volumes most analogous to the probability model;

reproducing contents wherein a content reproducing part reproduces contents in accordance with the playlist generated in the step of generating a playlist;

generating a content reproduction history wherein the content reproducing part generates and stores content reproduction history information in a content reproduction history information holding part; and

adapting a model wherein a model adapting part performs a process in which the model adapting part adapts a parameter included in a time series of content feature volumes corresponding to a reproduction history matched with content reproduction history information stored in the content reproduction history information holding part to a probability model held in the model information holding part, whereby the model adapting part approximates the probability model to the time series of content feature volumes corresponding to the reproduction history.

8. The information processing method according to claim 7, further comprising the steps of:

generating a content reproduction history wherein the content reproducing part generates and stores content reproduction history information in a content reproduction history information holding part; and

selecting an optimal model wherein an optimal model selecting part compares a time series pattern of content feature volumes corresponding to a reproduction history matched with content reproduction history information stored in the content reproduction history information holding part with a plurality of probability models held in the model information holding part, and selects a probability model most analogous to the time series pattern of content feature volumes corresponding to the reproduction history as an optimal model,

wherein the step of generating a playlist is the step of comparing a time series pattern of the content feature volumes extracted in the step of extracting a content feature corresponding to each of a plurality of permutation patterns of a reproduction sequence of contents to be reproduced with the optimal model, and generating a playlist in which a reproduction sequence of contents corresponding to a time series pattern of content feature volumes most analogous to the optimal model.

9. The information processing method according to claim 7, further comprising the step of:

generating a model wherein a model creating part generates a probability model to which sample data containing a reproduction sequence of contents is adapted, performs a learning process for the created probability model to create a plurality of different probability models, and stores the created probability model in the model information holding part.

10. The information processing method according to claim 7, further comprising the step of:

selecting contents to be reproduced wherein a reproduction content selecting part selects a plurality of contents configured of a time series pattern of content features volumes analogous to a probability model held in the model information holding part as contents to be reproduced.

11. The information processing method according to claim 7,

wherein the model information holding part is configured to store therein a probability model in accordance with a hidden Markov model (HMM) as a probability model corresponding to a time series pattern of content feature volumes, the content feature volume being feature information about a content.

12. The information processing method according to claim 7,

wherein the content feature volume includes at least any one of feature information about a data signal configuring a content and metainformation corresponding to a content.

13. A computer program which allows an information processing apparatus to perform a process for generating a playlist that defines a reproduction sequence of contents, the process comprising the steps of:

extracting a content feature wherein a content feature extracting part is allowed to acquire a content feature volume corresponding to each of a plurality of contents to be reproduced;

generating a playlist wherein a playlist generating part is allowed to compare a time series pattern of the content feature volumes extracted in the step of extracting a content feature corresponding to each of a plurality of permutation patterns of a reproduction sequence of contents to be reproduced with a probability model held in advance in a model information holding part and to generate a playlist in which a reproduction sequence of contents is set in accordance with a time series pattern of content feature volumes most analogous to the probability model;

reproducing contents wherein a content reproducing part is allowed to reproduce contents in accordance with the playlist generated in the step of generating a playlist;

generating a content reproduction history wherein the content reproducing part generates and stores content reproduction history information in a content reproduction history information holding part; and

adapting a model wherein a model adapting part performs a process in which the model adapting part adapts a parameter included in a time series of content feature volumes corresponding to a reproduction history matched with content reproduction history information stored in the content reproduction history information holding part to a probability model held in the model information holding part, whereby the model adapting part approximates the probability model to the time series of content feature volumes corresponding to the reproduction history.