en
en-es
0.25
0.5
0.75
1.25
1.5
1.75
2
Spacetime Forests with Complementary Features for Dynamic Scene Recognition
Published on 2014-04-032834 Views
This paper presents spacetime forests defined over complementary spatial and temporal features for recognition of naturally occurring dynamic scenes. The approach improves on the previous state-of-t
Related categories
Presentation
Spacetime Forests with Complementary Features for Dynamic Scene Recognition00:00
Dynamic Scene Classification00:07
Challenges (1)00:27
Challenges (2)00:54
Challenges (3)01:20
Challenges (4)01:39
Related work02:03
Proposed temporal slicing02:51
Approach overview03:34
Complementary Spacetime Orientation (CSO) descriptor: Temporal information - 104:57
Complementary Spacetime Orientation (CSO) descriptor: Temporal information - 206:15
Local contrast normalization06:31
Complementary Spacetime Orientation (CSO) descriptor: Spatial information07:31
Complementary Spacetime Orientation (CSO) descriptor: Temporal information08:04
Complementary Spacetime Orientation (CSO) descriptor: Colour information08:54
Spacetime energy pooling09:20
Random forest classifier - 109:33
Random forest classifier - 209:47
Random forest classifier - 309:53
Spacetime Random Forest (STRF)10:04
Classification with spacetime forest10:42
Maryland “in the wild” dataset11:20
Results on Maryland “in the wild”11:50
YUPENN dynamic scenes dataset13:15
Results on YUPENN dynamic scenes13:30
Complementarity of CSO descriptor14:19
In Summary15:10