This document presents a relevance feedback approach to video genre retrieval. It discusses using relevance feedback algorithms like Rocchio, feature relevance estimation, SVM, and hierarchical clustering relevance feedback to improve content-based video retrieval systems. An evaluation of these algorithms on a test database of over 90 hours of video across 7 genres showed that the hierarchical clustering relevance feedback approach outperformed other methods, achieving a mean precision improvement of 76.61% compared to 40.82% for the initial descriptor alone. Future work could involve testing the approach on larger, more challenging databases.
Roadmap to Membership of RICS - Pathways and Routes
Video Genre Retrieval Using Hierarchical Clustering Relevance Feedback
1. University Politehnica of Bucharest - Faculty of Electronics and Telecommunications
A Relevance Feedback Approach to
Video Genre Retrieval
Ionut MIRONICA1 Constantin VERTAN1 Bogdan IONESCU1,2
1LAPI – University Politehnica of Bucharest, Romania
52LISTIC – Université de Savoie, Annecy, 74944 France
IEEE International Conference on Intelligent Computer Communication
and Processing - Cluj Napoca – ICCP 2011
Sunday, November 2, 2014 ICCP 2011 1
2. University Politehnica of Bucharest - Faculty of Electronics and Telecommunications
Agenda
• Introduction
• Relevance Feedback Algorithms
• Hierarchical Clustering Relevance Feedback
• Implementation and Evaluation
• Conclusion
Sunday, November 2, 2014 ICCP 2011 2
3. University Politehnica of Bucharest - Faculty of Electronics and Telecommunications
I. Introduction
Concepts
• Content Based Video Retrieval
• Query by Example
Query Results
Query Database
Movie Sample
Sunday, November 2, 2014 ICCP 2011 3
4. University Politehnica of Bucharest - Faculty of Electronics and Telecommunications
II. CBVR System
Video
Descriptors
(vectors with
features)
Movie Database
(Off-line)
Descriptor
compute
(On-line)
Descriptor
Compute
Comparison
Image
Selection
(browsing)
Relevance feedback
Sunday, November 2, 2014 ICCP 2011 4
5. University Politehnica of Bucharest - Faculty of Electronics and Telecommunications
II. Color Descriptors
Objective: capture movie’s global color contents in terms of color
distribution, elementary hues, color properties and color relationship;
Global weighted histogram:
é
M
( ) 1 ( ) w
N
i
å å
= =
ù
× úû
= ×
êë
i
i
j
j
shot
0 0
i
h c
GW
i h c
N
B
A
c o u l e u r s c h a u d e s c o u l e u r s a d j a c e n t e s
c o u l e u r s f r o i d e s
c o u l e u r s c o m p l é m e n t a i r e s
Elementary color histogram:
215
Name(c ) Name(c) e Ì å==
Webmaster 216 colors Itten’s color wheel
[B. Ionescu, D. Coquin, P. Lambert, V. Buzuloiu’08]
E e GW h c h c
( ) ( )
0
c
Sunday, November 2, 2014 ICCP 2011 5
6. University Politehnica of Bucharest - Faculty of Electronics and Telecommunications
II. Color Descriptors
Color properties
= å
light GW P h c Ì
= : amount of bright colors in the movie,
light ( ) Name c
W ( )
215
0
c
WlightÎ{light, pale, white};
dark P : amount of dark colors in the movie, WdarkÎ{dark, obscure, black};
: amount of saturated colors, WhardÎ{hard, faded} È elem.;
: amount of low saturated colors, WweakÎ{weak, dull};
hard P
weak P
warm P
cold P
: amount of warm colors, WwarmÎ{Yellow, Orange, Red, Yellow-Orange,
Red-Orange, Red-Violet, Magenta, Pink, Spring};
: amount of cold colors, WcoldÎ{Green, Blue, Violet, Yellow-Green,
Blue-Green, Blue-Violet, Teal, Cyan, Azure};
Sunday, November 2, 2014 ICCP 2011 6
7. University Politehnica of Bucharest - Faculty of Electronics and Telecommunications
II. Action Descriptors
Objective: capture movie’s temporal structure in terms of visual rhythm,
action and gradual transition %;
Gradual transition %:
T T T
GT - - + +
dissolve fade in fade out
T
total
=
Rhythm: capture the movie’s changing tempo
( ) 5 i T = s x : relative number of shot changes within time window T
starting from frame at time index i;
{ ( )} v 5 E 5 i T = s = xT = s : movie’s average shot change speed;
Action: in general related to a high frequency of shot changes;
“hot action” “low action”
[B. Ionescu, L. Ott, P. Lambert, D. Coquin, A. Pacureanu, V. Buzuloiu’10]
Sunday, November 2, 2014 ICCP 2011 7
8. University Politehnica of Bucharest - Faculty of Electronics and Telecommunications
II. Contour Descriptors
Objective: describe structural information in terms of contours and
their relations;
Contour properties:
b
z : degree of circularity; – ½ circle vs. full circle
: degree of curvature (proportional to the maximum amplitude of the
bowness space); – straight vs. bow
edginess
symmetry
[IJCV, C. Rasche’10]
e : edginess parameter – zig-zag vs. sinusoid;
y : symmetry parameter – irregular vs. “even”
+ Appearance parameters:
m s: mean, std.dev. of intensity along the contour; c , c
m s: fuzziness, obtained from a blob (DOG) filter: I * DOG f , f
Sunday, November 2, 2014 ICCP 2011 8
9. University Politehnica of Bucharest - Faculty of Electronics and Telecommunications
III. Semantic Gap and Relevance Feedback
• Relevance feedback uses positive and negative
examples provided by the user to improve the
system’s performance.
User
Feedback
Display
Feedback
to system
Estimation &
Display selection
Display
User
Feedback
Sunday, November 2, 2014 ICCP 2011 9
10. University Politehnica of Bucharest - Faculty of Electronics and Telecommunications
III. Relevance Feedback
Algorithms:
• Rocchio Algorithm
• Feature Relevance Estimation (FRE)
• SVM – Relevance Feedback
• Hierarchical Clustering Relevance Feedback
Sunday, November 2, 2014 ICCP 2011 10
11. University Politehnica of Bucharest - Faculty of Electronics and Telecommunications
III. Rocchio Algorithm
• Uses a set R of relevant documents and a set N of
non-relevant documents, selected in the user relevance
feedback phase, and updates the query feature.
Q' aQ b g
å R
å
Î Î
= + -
N N
i
R R
i
N
i i
N
R
Sunday, November 2, 2014 ICCP 2011 11
12. University Politehnica of Bucharest - Faculty of Electronics and Telecommunications
III. Feature Relevance Estimation (FRE)
• some specific features may be more important than other
features
• every feature will have an importance weight that will be
computed as Wi = 1/σ, where σ denotes the variance of
video features.
d
d
å å
= =
( , ) ( )2
Dist X Y = W X -
Y W
i i i i i
i
1 1
Sunday, November 2, 2014 ICCP 2011 12
13. University Politehnica of Bucharest - Faculty of Electronics and Telecommunications
III. SVM – Relevance Feedback
• Train SVM with two classes:
- positive samples
- negative samples
• Classify next samples as positive or negative using
SVM network
• Use the confidence values of the classifiers to sort the
images
Sunday, November 2, 2014 ICCP 2011 13
14. University Politehnica of Bucharest - Faculty of Electronics and Telecommunications
IV. Hierarchical Clustering RF
Calculate the similarity between all
possible combinations of two videos
Two most similar clusters are
grouped together to form a new
cluster
Calculate the similarity between the
new cluster and all remaining
clusters.
Stop
Condition
Classify next samples
acording to dendogram
hierarchy
End
Sunday, November 2, 2014 ICCP 2011 14
15. University Politehnica of Bucharest - Faculty of Electronics and Telecommunications
IV. Hierarchical Clustering RF
Stop Condition
Variant 1:
- The number of clusters = a fixed number ( e.q. quarter of
the number of videos within a retrieved batch)
Variant 2:
- The number of clusters = an adaptive number
D
=1- where Dij represents the distance between
ij
D
ij
d
min
max
two clusters
S2u5n.0d5a.0y9, November 2, 2014 ICCP 2011 15
16. University Politehnica of Bucharest - Faculty of Electronics and Telecommunications
IV. Hierarchical Clustering RF
Similarity measures – Centroid Distance
+
+
C1
C2
Centroid
Centroid
S2u5n.0d5a.0y9, November 2, 2014 ICCP 2011 16
17. University Politehnica of Bucharest - Faculty of Electronics and Telecommunications
IV. Hierarchical Clustering RF
How will be clustered?
Initial
Query
Sunday, November 2, 2014 ICCP 2011 17
18. University Politehnica of Bucharest - Faculty of Electronics and Telecommunications
IV. Hierarchical Clustering RF
Sunday, November 2, 2014 ICCP 2011 18
19. University Politehnica of Bucharest - Faculty of Electronics and Telecommunications
IV. Hierarchical Clustering RF
Sunday, November 2, 2014 ICCP 2011 19
20. University Politehnica of Bucharest - Faculty of Electronics and Telecommunications
IV. Hierarchical Clustering RF
Sunday, November 2, 2014 ICCP 2011 20
21. University Politehnica of Bucharest - Faculty of Electronics and Telecommunications
V. Implementation and Evaluation
The Test database
- 91 hours of video
• 20h30m of animated movies (long, short clips and
series),
• 15m of TV commercials,
• 22h of documentaries (wildlife, ocean, cities and
history),
• 21h57m of movies (long, episodes and sitcom),
• 2h30m of music (pop, rock and dance video clips),
• 22h of news broadcast
• 1h55min of sports (mainly soccer) (a total of 210
sequences, 30 per genre).
Sunday, November 2, 2014 ICCP 2011 21
22. University Politehnica of Bucharest - Faculty of Electronics and Telecommunications
V. Implementation and Evaluation
• Precision – Recall Chart
AP d precision
Sunday, November 2, 2014 ICCP 2011
22
Feedback Approaches for MPEG-7 Content-based Biomedical Image
Precision =
Number of returned relevant images
Total number of returned images
Recall =
Number of returned relevant images
Total number of relevant images
• Average Precision
å=
=
d
i
1
1/
23. University Politehnica of Bucharest - Faculty of Electronics and Telecommunications
V. Implementation and Evaluation (1)
Evaluation for the first feedback
Precision-recall curves per genre category
curves for different descriptors:
- Color & Action
- Contour
- Color & Action & Contour
Sunday, November 2, 2014 ICCP 2011 23
24. University Politehnica of Bucharest - Faculty of Electronics and Telecommunications
V. Implementation and Evaluation (2)
Evaluation per video genre
Initial Descriptor 40.82%
Rocchio 58.20%
Robertson/Starck-Jones 55.83%
Feature Relevance Estimation 68.48%
Support Vector Machines 70.28%
Hierarchical Clustering RF 76.61%
Mean precision improvement with Relevance Feedback
Sunday, November 2, 2014 ICCP 2011 24
25. University Politehnica of Bucharest - Faculty of Electronics and Telecommunications
VI. Conclusions
• Relevance feedback is a powerful tool for improving
content based video retrieval systems
• The Hierarchical Clustering RF Approach outperforms
classical RF algorithms (such as Rocchio or RFE and
SVM) in terms of accuracy and computational effort.
Future Work
- Test the algorithm on bigger and more difficult databases
with more classes, e.g. MediaEval 2011 – 26 genres, 2500
sequences from Blip.Tv (social media platform)
Sunday, November 2, 2014 ICCP 2011 25
26. University Politehnica of Bucharest - Faculty of Electronics and Telecommunications
Thank you!
Questions?
Sunday, November 2, 2014 ICCP 2011 26