U.S. patent application number 12/457429 was filed with the patent office on 2010-08-05 for dynamic video segment recommendation based on video playback location.
This patent application is currently assigned to Napo Enterprises. Invention is credited to Scott Curtis, Mike Helpingstine, Kunal Kandekar, Ravi Reddy Katpelly.
Application Number | 20100199295 12/457429 |
Document ID | / |
Family ID | 42398515 |
Filed Date | 2010-08-05 |
United States Patent
Application |
20100199295 |
Kind Code |
A1 |
Katpelly; Ravi Reddy ; et
al. |
August 5, 2010 |
Dynamic video segment recommendation based on video playback
location
Abstract
A system and method for dynamically providing video segment
recommendations to a user based on video metadata, current playback
position within the video, and/or the user's previous viewing
history or patterns. The system/method also provides the user with
a simple mechanism to enable/disable receipt of these
recommendations with the click of a button within the display
interface.
Inventors: |
Katpelly; Ravi Reddy;
(Durham, NC) ; Kandekar; Kunal; (Jersey City,
NJ) ; Helpingstine; Mike; (Chapel Hill, NC) ;
Curtis; Scott; (Durham, NC) |
Correspondence
Address: |
AKERMAN SENTERFITT
8100 BOONE BOULEVARD, SUITE 700
VIENNA
VA
22182-2683
US
|
Assignee: |
Napo Enterprises
Wilmington
DE
|
Family ID: |
42398515 |
Appl. No.: |
12/457429 |
Filed: |
June 10, 2009 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
61149216 |
Feb 2, 2009 |
|
|
|
Current U.S.
Class: |
725/14 ;
725/46 |
Current CPC
Class: |
G06N 5/02 20130101; G06F
16/635 20190101; H04L 67/306 20130101; H04L 65/60 20130101; G06F
16/4387 20190101; G06F 16/686 20190101; G06N 20/00 20190101; G06N
5/04 20130101 |
Class at
Publication: |
725/14 ;
725/46 |
International
Class: |
H04H 60/32 20080101
H04H060/32; H04N 5/445 20060101 H04N005/445 |
Claims
1. A method of dynamically recommending media segments to a user,
comprising: analyzing and gathering metadata of a currently viewed
media segment of a main media content based on a current playback
location; identifying other media segments, from the main media
content or other media content, that are similar to the currently
viewed media segment; and dynamically recommending at least some of
the identified other, similar media segments to the user.
2. The method of claim 1, wherein the media segments comprise video
segments.
3. The method of claim 2, wherein only after the currently viewed
video segment has been played back for a time period greater than a
preset time period threshold are other, similar video segments
identified.
4. The method of claim 2, wherein the metadata is identified as at
least one of characters, speech, audio effects, soundtrack,
semantics, events in a scene, timing, or a scene location.
5. The method of claim 4, further comprising: matching the metadata
of the currently viewed video segment against metadata of all video
segments available in a video database; and storing a list of all
the video segments that match.
6. The method of claim 2, further comprising: gathering user
information including one of a previous viewing history of the
user, a user profile, or a user preference; and finding video
segments identified that most closely match with the gathered user
information.
7. The method of claim 6, further comprising: identifying the found
video segments that have been determined to match one of a viewing
history, profile, or preference of other users.
8. The method of claim 6, further comprising: displaying the video
segments that most closely match with the previous viewing history
of the user.
9. The method of claim 8, wherein the video segments are displayed
as at least one of video clips, links, pop-ups or
notifications.
10. The method of claim 2, wherein the recommended, similar video
segments are displayed upon request by the user.
11. The method of claim 2, wherein the recommended, similar video
segments are displayed upon subsequent video segment changes.
12. The method of claim 1, further comprising continuously updating
the similar media segments depending on the current playback
location of the user.
13. The method of claim 1, further comprising clicking a button on
a display interface to permit the user to enable/disable receipt of
the recommended, similar media segments.
14. A video scene similarity assessment method, comprising:
analyzing and gathering metadata of a currently viewed video scene
of a main video content based on a current playback location;
matching the metadata of the currently viewed video scene against
metadata of all video scenes available in a collection of videos,
the metadata being identified as at least one of characters,
speech, audio effects, soundtrack, semantics, events in a scene,
timing, or a scene location; and storing a list of all the video
scenes that match.
15. The method of claim 14, further comprising: identifying other
video scenes, from the main video content or other video content,
that are similar to the currently viewed video scene; gathering
user information including one of a previous viewing history of a
user, a user profile, or a user preference; and finding video
scenes identified that most closely match with the gathered user
information.
16. A media system for dynamically recommending video scenes to a
user, comprising: means for detecting metadata of a currently
viewed video scene of a video content; means for identifying other
video scenes, from the video content or other video content, that
are similar to the currently viewed video scene; and means for
dynamically recommending at least some of the identified other,
similar video scenes to the user based on a previous viewing
history of the user.
17. The media system of claim 16, wherein the metadata is detected
based on a current playback location.
18. The media system of claim 16, wherein only after the currently
viewed video scene has been played back for a time period greater
than a preset time period threshold are other, similar video scenes
identified.
19. The media system of claim 16, wherein the metadata is
identified as at least one of characters, speech, audio effects,
soundtrack, semantics, events in a scene, timing, or a scene
location.
20. The media system of claim 19, further comprising: means for
matching the metadata of the currently viewed video scene against
metadata of all video scenes available in a video database; and
means for storing a list of all the video scenes that match.
21. The media system of claim 16, further comprising: means for
gathering user information including one of a previous viewing
history of the user, a user profile, or a user preference; and
means for finding video scenes identified that most closely match
with the gathered user information.
22. The media system of claim 21, further comprising: means for
identifying the found video scenes that have been determined to
match one of a viewing history, profile, or preference of other
users.
23. The media system of claim 22, wherein the other users include
users from a social network of the user, or users that have a
profile similar to the user, or users whose previous video uploads
are similar in characteristics to video uploads of the user, or
users who match in terms of type of videos that the user
watches.
24. The media system of claim 21, further comprising: means for
displaying the video scenes that most closely match with the
previous viewing history of the user.
25. The media system of claim 24, wherein the video scenes are
displayed as at least one of video clips, links, pop-ups or
notifications.
26. The media system of claim 16, wherein the recommended, similar
video scenes are displayed upon request by the user.
27. The media system of claim 16, wherein the recommended, similar
video scenes are displayed upon subsequent video scene changes.
28. The media system of claim 17, wherein the means for dynamically
recommending continuously updates the recommended, similar video
scenes depending on the current playback location of the user.
29. The media system of claim 16, further comprising means for
enabling/disabling receipt of the recommended, similar video
scenes.
30. A computer readable medium comprising software for instructing
a media system to: detect and gather metadata of a currently viewed
media segment of a primary media content based on a current
playback location; identify other media segments, from the primary
media content or additional media content, that are similar to the
currently viewed media segment; and dynamically recommend at least
some of the identified additional, similar media segments to a user
based on previous viewing patterns of the user.
31. A media system for dynamically recommending video segments to a
user, comprising: a media player which detects metadata of a
currently viewed video scene of a video content based on a current
playback location within the video content; a segment similarity
analyzer which identifies other video scenes, from the video
content or other video content, that are similar to the currently
viewed video scene, and dynamically recommends at least some of the
identified other, similar video scenes to the user based on a
previous viewing history of the user; and a display device which
displays the recommended, similar video scenes that most closely
match with the previous viewing history of the user.
32. The media system of claim 31, wherein the recommended, similar
video scenes are displayed as at least one of video clips, links,
pop-ups or notifications.
33. The media system of claim 31, wherein the media system
continuously updates the recommended, similar video scenes
depending on the current playback location of the user.
34. The media system of claim 31, further comprising a button
disposed on a display interface of the display device which permits
the user to enable/disable receipt of the recommended, similar
video scenes, by clicking thereon.
35. The media system of claim 31, wherein only after the currently
viewed video scene has been played back for a time period greater
than a preset time period threshold are other, similar video scenes
identified.
36. The media system of claim 31, wherein other, similar video
scenes are identified and one of selected, not selected, or
prioritized based on whether a trick play mode, including at least
one of a skip mode or a fast forward mode, has taken place with
respect to the currently viewed video scene.
37. The method of claim 7, wherein the other users include users
from a social network of the user, or users that have a profile
similar to the user, or users whose previous video uploads are
similar in characteristics to video uploads of the user, or users
who match in terms of type of videos that the user watches.
Description
CROSS REFERENCE TO RELATED APPLICATIONS
[0001] The present application claims priority from U.S.
Provisional Application No. 61/149,216 filed on Feb. 2, 2009, the
disclosure of which is incorporated herein by reference in its
entirety.
FIELD OF THE INVENTION
[0002] The present invention relates to a media system and, more
particularly, to a media system for and method of dynamically and
intelligently recommending video segments to a user.
BACKGROUND OF THE INVENTION
[0003] In general, video sharing sites such as YouTube allow users
to browse/search for videos, and view the videos as and when they
find anything interesting. Users can even browse through
similar/related videos while watching a certain video.
[0004] However, if a user likes a certain segment/scene in a video
and would like to find/view any similar scenes, current systems
provide no search feature (to find "Related Video Segments"). In
such cases, users are required to recreate and re-submit a new
search query, which may not be easy for users to formulate (or
describe in words) especially when a user is interested in a
particular scene.
[0005] More specifically, one related art system that enables users
to browse through similar/related videos utilizes an MPEG-7 based
metadata structure for associating/grouping video segments with
similar/related events taking place. The events can be, for
example, all touchdown passes in a football game, etc. These events
may be part of the same video or may belong to different
videos.
[0006] Several extensions have been proposed to the model. For
example, a sports reporter who is trying to create a reportage on
touch down passes in a football game may look for all the scenes
where touchdowns took place, and associate them together as a
"touchdown" event. As an another example, for creating a reportage
on "touchdown" events from several games by a single player (e.g.,
Peyton Manning), the reporter can identify a "touchdown" event and
a "Manning" object, and find all scenes where Manning threw
touchdown passes, and associate all such events found to the
"Manning" object.
[0007] The related art system offers several extensions to the
model to provide a lean-back viewing experience for an end-user.
However, the users are required to tell the system via a preference
profile page, of the events that they are interested in and would
like to receive the recommendations. For example, a user may
indicate his preference to "touchdown" events, and explicitly
mention in his preference profile that he would like to receive
"touchdown" events either at the end of the event or at the end of
the video. With this model, users are required to explicitly
subscribe to the system of desired events. Next, while watching
another or the same video, if the user likes another event, he is
required to again update his preference profile of that event.
[0008] Furthermore, describing/representing an event in plain text
itself might not be intuitive to the users. For example, say if the
user is watching a football game and is interested in a "goal"
event (meaning that while the football is near the goal line), such
events cannot be easily represented in a way that the
recommendation engine can interpret it correctly. The "goal" event
can be understood as "goals scored in a soccer game". In such a
case, the recommendation engine would supply video segments
containing "goals scored in a soccer game".
SUMMARY OF THE INVENTION
[0009] The present invention provides a system that can dynamically
show similar segments to a user based on the current playback
location, and, for example, based on the user's previous viewing
history (patterns).
[0010] The present invention makes recommendations that are
implicit in nature, i.e., users are not required to explicitly
indicate their interest in certain events. These recommendations
are based on the current playback location of the user in the video
and utilize a user's profile, preferences and previous viewing
history behavior to determine which of the recommendations best
suit the user's interest. Unlike related art systems where a user's
preference profile is used by the system to check if the user is
interested in a given video, or the current event (within the
video), the present invention uses a user's preference profile to
select the best of all the available (or dynamically generated)
recommendations that would suit the user's interest with respect to
current activity (or event) in the video. The present invention can
therefore provide continuous video segment recommendations to the
user (i.e., as the current segment being viewed keeps changing, the
recommendations keep changing).
[0011] The present invention provides a system and method that
identifies and displays video segments that are similar to a user's
currently viewed video segment. The system dynamically and
continuously updates the similar segments depending on the user's
current playback position. As and when the video progresses or
depending on the user's playback events (such as skip, fast
forward, rewind, etc), the similar segments are updated. The
similar segments can be displayed on user request (such as when a
user clicks on video player to view similar segments), or can be
displayed/updated periodically (such as every ten seconds), or can
be displayed/updated based on changes or events in the viewed video
(such as at the beginning of a new scene). The users may optionally
specify how frequently they wish to receive the recommendations, at
what point within the scene (i.e., the offset from the start of the
scene) they may want to receive them, and so forth. The
system/method can be used in any distributed as well as centralized
system, and can be used as a notification service to receive video
scene recommendations.
[0012] According to one aspect, the present invention provides a
method of dynamically recommending media segments to a user,
comprising: analyzing and gathering metadata of a currently viewed
media segment of a main media content based on a current playback
location; identifying other media segments, from the main media
content or other media content, that are similar to the currently
viewed media segment; and dynamically recommending at least some of
the identified other, similar media segments to the user.
[0013] According to another aspect, the media segments may comprise
video segments.
[0014] According to another aspect, the method may further comprise
matching the metadata of the currently viewed video segment against
metadata of all video segments available in a collection of videos
such as a video database; and storing a list of all the video
segments that match. The database also may be a centralized
database or a distributed database or a peer-to-peer (P2P) system.
Thus, a query may be issued containing the current segment metadata
across a distributed system to find a list of matching segments.
Also, the database may store video metadata instead of the videos,
and the metadata may include a pointer to the actual videos.
[0015] According to another aspect, the method may further comprise
gathering user information including one of a previous viewing
history of a user, a user profile, or a user preference; and
finding video segments identified that most closely match with the
gathered user information. The user's previous browsing history may
also be used to filter matching segments, such as, for example,
those segments that the user has already watched.
[0016] According to another aspect of the present invention, a
media system for dynamically recommending video scenes to a user is
provided, comprising: means for detecting metadata of a currently
viewed video scene of a video content; means for identifying other
video scenes, from the video content or other video content, that
are similar to the currently viewed video scene; and means for
dynamically recommending at least some of the identified other,
similar video scenes to the user based on the user's previous
viewing history.
[0017] The present invention also contemplates a computer readable
medium comprising software for instructing a media system to:
detect and gather metadata of a currently viewed media segment of a
primary media content based on a current playback location;
identify other media segments, from the primary media content or
additional media content, that are similar to the currently viewed
media segment; and dynamically recommend at least some of the
identified additional, similar media segments to a user based on
previous viewing patterns of the user.
[0018] A media system for dynamically recommending video segments
to a user, comprising: a media player which detects metadata of a
currently viewed video scene of a video content based on a current
playback location within the video content; a segment similarity
analyzer which identifies other video scenes, from the video
content or other video content, that are similar to the currently
viewed video scene, and dynamically recommends at least some of the
identified other, similar video scenes to the user based on a
previous viewing history of the user; and a display device which
displays the recommended, similar video scenes that most closely
match with the previous viewing history of the user.
[0019] Those skilled in the art will appreciate the scope of the
present invention and realize additional aspects thereof after
reading the following detailed description of the preferred
embodiments in association with the accompanying drawing
figures.
BRIEF DESCRIPTION OF THE DRAWING FIGURES
[0020] The accompanying drawing figures incorporated in and forming
a part of this specification illustrate several aspects of the
invention, and together with the description serve to explain the
principles of the invention.
[0021] FIG. 1 illustrates a media system according to an exemplary
embodiment of the present invention;
[0022] FIG. 2 depicts an illustrative embodiment of a method
operating in the media system of FIG. 1;
[0023] FIG. 3 depicts a further illustrative embodiment of a method
operating in the media system of FIG. 1;
[0024] FIG. 4 depicts a still further illustrative embodiment of a
method operating in the media system of FIG. 1;
[0025] FIGS. 5A and 5B illustrate examples of the operation of the
present invention; and
[0026] FIG. 6 illustrates a further example of the operation of the
present invention.
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0027] The embodiments set forth below represent the necessary
information to enable those skilled in the art to practice the
invention. Upon reading the following description in light of the
accompanying drawing figures, those skilled in the art will
understand the concepts of the invention and will recognize
applications of these concepts not particularly addressed herein.
It should be understood that these concepts and applications fall
within the scope of the disclosure and the accompanying claims.
[0028] Note that at times the system of the present invention is
described as performing a certain function. However, one of
ordinary skill in the art would know that the program is what is
performing the function rather than the entity of the system
itself.
[0029] Although aspects of one implementation of the present
invention are depicted as being stored in memory, one skilled in
the art will appreciate that all or part of systems and methods
consistent with the present invention may be stored on or read from
other computer-readable media, such as secondary storage devices,
like hard disks, floppy disks, and CD-ROM, a carrier wave received
from a network such as the Internet, or other forms of ROM or RAM
either currently known or later developed. Further, although
specific components of the system have been described, one skilled
in the art will appreciate that a system suitable for use with the
methods and systems consistent with the present invention may
contain additional or different components.
[0030] FIG. 1 illustrates a media system according to an exemplary
embodiment of the present invention. In general, the media system
for dynamic video segment recommendations 10 includes an input
device 12, such as but not limited to a keyboard, keypad,
smartphone, or remote control for operation by an associated user
14, and a media playback system 16. In this exemplary embodiment,
the media playback system 16 includes a media player 18 and a
display device 20.
[0031] The media player 18 may be, for example, a personal
computer, a set-top box (STB) for playing digital television
content received from a television content provider, a Digital
Video Recorder (DVR) for playing previously recorded video content
such as previously recorded television content received from a
television content provider, an Apple TV.RTM. device for playing
downloaded content that has been purchased or rented from a remote
media distribution service such as the Apple.RTM.D iTunes.RTM.
store, a Digital Versatile Disc (DVD) player, or the like. The
media player 18 may be connected to the display device 20 via any
desired audio/video connection such as, for example, a High
Definition Multimedia Interface (HDMI) connection, a Digital Video
Interface (DVI) connection, a coaxial cable connection, or the
like. The display device 20 may be, for example, a computer display
screen, a television (TV), or the like. In an alternative
embodiment, the display device 20 may be incorporated into the
media player 18.
[0032] The media player 18 includes a media playback function 24
and a media segment playback function 26, each of which may be
implemented in software, hardware, or a combination thereof. The
media playback function 24 generally operates to provide playback
of media items obtained from a content source 28. In the exemplary
embodiment, the media items are video items. As such, the media
playback function 24 provides playback of the video items and
presentation of the video items to the user 14 and any other nearby
users via the display device 20. The content source 28 varies
depending on the particular implementation of the media player 18.
For example, if the media player 18 is a STB, then the content
source 28 may be a television content distribution network such as
a Cable Television (CATV) network. As another example, if the media
player 18 is a DVD player, then the content source 28 is a DVD. As
a final example, if the media player 18 is a device such as an
Apple TV.RTM. device, then the content source 28 may be a remote
media distribution service such as the Apple.RTM. iTunes.RTM.
store, where the media player 18 has access to the remote media
distribution service via a network such as, for example, the
Internet.
[0033] With reference to FIGS. 1 and 2, in one exemplary embodiment
of the present invention, while a user 14 is watching a video (step
S100), the media player 18 notifies a Segment Similarity Analyzer
(SSA) 30 of the current segment the user is viewing (step S102).
Along with this is sent information such as user's current playback
location, metadata information associated to that segment, metadata
of the video scene for that segment, and so forth. Once the SSA 30
receives the information, it looks for video scenes that have
metadata similar to the current segment/scene from, for example, a
video database (step S104). The video database may be part of the
SSA 30, or an external source. The metadata similarities can be
identified in terms of characters, speech, semantics, events in the
scene, timing, scene location, audio effects, soundtrack and so
forth. Any model that uses MPEG-7 framework for associating
metadata similarities among videos can be used. An example of one
such model is disclosed in the related art document entitled, "A
Video Metadata Model Supporting Personalization &
Recommendation in Video-based Services" by Tsinaraki et al. (July
2001), the disclosure of which is incorporated herein by reference
in its entirety.
[0034] Moreover, for each segment, the metadata information
identifying and describing the segment may include information
describing the content of the segment of the media item. For
example, the information may describe the segment as containing an
action scene, a romantic scene, or the like. As another example, if
the media item is one of the Star Wars movies, the information may
describe the content of the segment more specifically as containing
a Princess Leia scene, a Darth Vader scene, a droid scene, a
space-fighting scene, or the like. As another example, the
information describing the segment may include a list of actors or
actresses appearing in the segment and/or a description of
activities that take place in the segment. The information
describing the content of the segments of the media item may be
information provided by a producer or creator of the media item,
information such as annotations provided by one or more users that
have previously viewed the media item, or the like, or any
combination thereof.
[0035] Also consistent with the present invention, the metadata may
be, for example, tags, annotations, a script or lyrics for the
media item, closed-captioning information, sub-titles, or the like.
Moreover, the media segment playback function 26 may also utilize a
combination of audio and frame analysis techniques. For example, in
addition to using frame analysis techniques to detect violent
content such as, for example, smoke or blood pixels, the system may
also utilize audio analysis techniques for detecting audio content
such as in the form of, for example, gunshot sounds.
[0036] Once a matching video scene is found, the SSA 30 adds that
scene to its list of matches found (step S106). Note that this can
be a one-time process, and does not need to be repeated for every
other user. However, there are scenarios where it may be desired to
repeat this process. One example is a distributed scenario, where
each device has its own SSA 30 which performs this operation for
its own user. Another example is where new content is being
continually added to the video database, and newer results may be
wanted in the recommendations. For example, in a database or search
engine, the video metadata may be indexed, and the corresponding
scenes are then associated with the appropriate entry into the
index. Thus, segments from new videos would be appended to the
appropriate entries in the index. For a distributed system or a P2P
system, the process may need to be repeated multiple times.
[0037] Next, the system decides as to which of the scenes are to be
displayed to the user depending on a user's profile, preferences
and viewing history. The system collects metadata of the scenes
that the user had not preferred to watch or had skipped earlier. In
addition, the system may also collect metadata of the scenes that
the user has already watched, and hence need not be recommended
again. For instance, the system recommends a segment from the movie
"Rambo 4" when a user is watching the movie "UHF", and the user
then watches that segment. However, the same segment from "Rambo 4"
shows up as a potential recommendation for a segment in another
movie "Hot Shots 2". The system notices that the user has already
watched the recommended segment before, so it does not recommend it
again. Alternatively, it may be programmed to do the opposite and
recommend the segment from "Rambo 4" as well as the segment from
"UHF" (discovered from the user's viewing history) for that segment
in "Hot Shots 2". Any/all of such scenes are filtered out at this
stage. Also, scenes that the user has already watched may be
filtered out. Then, from among the rest of the scenes that survive
and match with the user's profile, the system identifies the scenes
that have been rated highly by other users or recommended by other
users. For example, this may include users from a user's social
network, or users that have a profile similar to the user, or the
users whose previous video uploads are similar in characteristics
to the user's video uploads, or users who match in terms of the
type of videos that the user watches. Finally, the scenes are
displayed to the user by the media segment playback function 26 of
the media device 18 in the order that best matched and/or were
rated highly (step S108). For example, if 100 scenes match a video
segment, a user X may see a scene 10 as the first matching video
scene, whereas a user Y may see a different scene 50 as his first
match. Note that the user can at any time opt not to see the
"Similar Scenes", by turning off a "Show Similar Scenes" button
(not shown).
[0038] Once the scenes to be displayed are identified, the system
can display them to the user in several ways, such as video clips,
links to clips, or pop-ups. The recommendations may also be
provided to a supplementary mobile device (e.g., a smartphone or a
remote control equipped with a display).
[0039] In step S110, if the main video is over, the user can end
the session, or if the user has viewed one of the matching scenes
or clips, the user can then return to the main video and continue
to watch it until completion.
[0040] Consider, for example, that a user is watching a football
game on Youtube. When the user reaches a video segment with
metadata identified as "touchdown", other segments with metadata
that includes "touchdown" are discovered and filtered based on his
previous history, and displayed to the user as recommendations, for
instance in the form of keyframes in the right-end of the browser
for viewing. The user at any point may continue to watch the
"touchdown" video segment from the current video or from the
recommendations made, i.e., the user is not required to wait until
the end of the event or the video to receive recommendations and
the recommendations will be updated as soon as his playback
position and the associated metadata changes (e.g., the current
metadata changes to "coach swearing" and the recommendation
keyframes get updated to show other instances of the same (or
other) coaches losing it).
[0041] With reference to FIG. 3, in another exemplary embodiment of
the present invention, the media player 18 may decide as to when
(or not) to notify the SSA 30 of the current segment (that the user
is watching (step S200)) based on some threshold parameters. For
example, the media player 18 may take into consideration the user's
trick play mode behavior (such as skip, fast forward, etc.) on the
current scene (step S202). Certain trick play modes on a scene are
indicative of the user finding the scene undesirable (e.g., user
skips the scene), and hence no segment information is sent to SSA
30 for scene analysis. For example, if the user 14 skips over a
particular segment of the media item, the media player 18 may
ensure that similar segments of the media item are not selected as
segments of interest during trick play mode, or may reduce the
priority assigned to similar segments. As another example, the
media player 18 may wait for the user to watch for a certain time
period within the scene (e.g., half the length of the scene (step
S204)), before which it notifies SSA 30 of the scene and playback
information. Thus, if the preset time period threshold has been
met, the media player 18 sends the SSA 30 the metadata of the scene
(step S206). The SSA 30 then analyzes the user profile and displays
recommendations to the user back at the media player 18 (step
S208).
[0042] With reference to FIG. 4, in another exemplary embodiment of
the present invention, the recommendations can be sent to the user
based on user preference of how and/or when he wants to receive
them. Such preferences can be indicated via the preference profile
page of the user with the media player 18. The media player 18 may
include a simple option (in a form of a button) as "Enable/disable
recommendations", and the user may simply click on the button to
turn the recommendations on/off (see, for example, button 60
discussed below). While the user is watching a video segment (step
S300) and selects to enable (to receive) the recommendations (step
S302), the media player 18 may look through user's preference as to
when and at what frequency he may wish to receive them (step S304).
For example, a user may wish to have the recommendations "only once
at the start of a new scene". Several other options include "user
can specify the frequency rate for receiving recommendations (e.g.,
twice within a scene) and within what part of the scene (such as
start/end/certain offset from the start of the scene, etc.)". If no
preferences are set by the user, a default action can be to send
recommendations at the beginning of every new scene.
[0043] The SSA 30 then analyzes the user profile for matching
recommendations and displays recommendations to the user at the
media player 18 (step S306).
[0044] Referring to FIG. 5A, consider, for example, that a user
(Bob) admires Barack Obama and is searching for "Barack Obama"
videos on his favorite online video service. The media system 10
returns a set of videos and Bob chooses one that looks interesting
and he watches it. While watching the video he comes across a
particular segment of the video that he enjoys--it features Mr.
Obama discussing the energy policy with his opponent John McCain.
Bob pauses the video on this segment 50 and then clicks on a "Find
Similar Segments" button 60 in the display of the media player 18.
In this case, Bob has indicated that recommendations be sent to him
only upon his request.
[0045] Clicking the button 60 instructs the SSA 30 of the media
system 10 to find other segments or scenes that contain content
that is similar to the current segment-content that is related by
metadata and/or by Bob's previous viewing patterns. For example,
the metadata may be in the form of segments containing key words
such as "Obama", "McCain" and "energy policy" thereby showing a
related segment of Mr. Obama. Alternatively, Bob's previous viewing
patterns could be based on Bob watching a particular Obama segment
five times. As shown in FIG. 5B, the media system 10 quickly
returns several segments 70A, 70B, 70C that have been bookmarked
for the appropriate segment to be queued for easy access by the
user.
[0046] Because the media system 10 has "watched" his behavior in
previous viewings of videos, it has determined that Bob is more
likely to prefer segments in which Obama is portrayed favorably.
Thus, it usually rules out segments provided by Fox News and favors
segments from MSNBC. Bob watches a returned segment (a segment
"recommended" by the system based on tags and Bob's preferences)
that features Obama speaking about "Change" in Germany. Bob again
pauses the player and clicks on "Find Similar Segments" button 60.
The media system 10 immediately returns a set of segments that are
uncannily similar to the speech in Germany segment.
[0047] With reference to FIG. 6, as an another example, consider
that a user (Bob) is interested in watching "Dog Show" videos on
Youtube. He has, however, explicitly indicated in his preferences
profile page that he would like to receive video segment
recommendations for any scenes that he has watched for at least 1
minute, and that the recommendations be presented to him only once
during that scene. He finds some videos on Youtube, but is
particularly interested in a video and starts watching it. He
however starts to skip some of the scenes in that video and comes
across a segment 80 with "Dogs Jumping" all around. Bob starts to
watch it and just when he is finished watching the first 1 minute
of that segment, he immediately finds a collection of scenes 90A,
90B, 90C, 90D that contain "Dog Jumping Clips" displayed to him (as
shown in FIG. 6 at the lower right).
[0048] He quickly scrolls to one of these clips, but this time Bob
would like to receive recommendations after he has watched a scene
for just 30 seconds. He would also like to continue to receive
recommendations repetitively at a fixed interval of 15 seconds from
that point on. He makes these changes in his preference profile
page, and starts to watch the clip. He comes across a segment with
"Boxer Hopping", and after watching the first 30 seconds of that
segment, he is presented with a collection of "Boxer Hopping
clips". None of those clips interests him, and after 15 seconds, he
next receives a new set of recommendations. As he finds them
interesting, he continues to watch the rest of the clips.
[0049] The present invention has substantial opportunity for
variation without departing from the spirit or scope of the present
invention. For example, while the embodiments discussed herein are
directed to personal or in-home playback, the present invention is
not limited thereto. Further, while the examples refer to video
segments or scenes, the present invention is not limited thereto
and other forms of media content are contemplated herein.
[0050] Those skilled in the art will recognize improvements and
modifications to the preferred embodiments of the present
invention. All such improvements and modifications are considered
within the scope of the concepts disclosed herein and the claims
that follow.
* * * * *