Selective max-pooling for object detection
View Patent ↗Systems and methods are disclosed for object detection by receiving an image and extracting features therefrom; applying a learning process to determine sub-regions and select predetermined pooling regions; and performing selective max-pooling to choose one or more feature regions without noises.
1. A method for object detection, comprising:
receiving an image and extracting features therefrom;
applying a learning process to determine sub-regions and select predetermined pooling regions;
performing selective max-pooling to choose one or more feature regions without noises,
forming at least an object bounding box for a location;
applying a cascaded boosting classifier to each object bounding box, with each weak classifier taking a feature response of a region inside the bounding box as its input and then the region is in tum represented by a group of small sub-regions (regionlets), and
determining a permutation invariant feature operation on features extracted from regionlets as
T
(
R
)
=
∑
j
=
1
N
R
α
j
T
(
r
j
)
,
subject
to
α
j
∈
{
0
,
1
}
,
∑
j
=
1
N
R
α
j
=
1
where T(R) as a feature representation for region R,T(r j ) as a feature extracted from the j th regionlet r j in R, N R is a total number of regionlets in region R, α j is a binary variable, either 0 or 1.
2. The method of claim 1 , comprising counting features appearing in candidate sub-regions.
3. The method of claim 1 , comprising selecting sets ofregionlets from a pool of candidate regionlet groups by boosting.
4. The method of claim 1 , comprising using a regionlet representation to model fine-grained spatial appearance layouts.
5. The method of claim 1 , comprising aggregating feature responses of regionlets within one group to a one dimensional feature robust to local deformation.
6. The method of claim 1 , comprising applying max-pooling over regionlet features as:
T
(
R
)
=
max
j
T
(
r
j
)
.
7. The method of claim 1 , comprising normalizing regionlets using detection windows.
8. The method of claim 1 , comprising normalizing regionlets by by using relative positions and sizes of the regionlets and their groups to an object bounding box.
9. The method of claim 1 , comprising training with boosting regionlet features.
10. The method of claim 1 , comprising using RealBoost to train cascaded classifiers for object detection.
11. A system for object detection, comprising:
a camera;
a processor coupled to the camera;
code for receiving an image and extracting features therefrom;
applying a learning process to determine sub-regions and select predetermined pooling regions; and
performing selective max-pooling to choose one or more feature regions without noises,
forming at least an object bounding box for a location;
applying a cascaded boosting classifier to each object bounding box, with each weak classifier taking a feature response of a region inside the bounding box as its input and then the region is in tum represented by a group of small sub-regions (regionlets), and
determining a permutation invariant feature operation on features extracted from regionlets as
T
(
R
)
=
∑
j
=
1
N
R
α
j
T
(
r
j
)
,
subject
to
α
j
∈
{
0
,
1
}
,
∑
j
=
1
N
R
α
j
=
1
where T(R) as a feature representation for region R,T(r j ) as a feature extracted from the j th regionletr j R, N R is a total number of regionlets in region R, α j is a binary variable, either 0 or 1.
12. The system of claim 11 , comprising code for counting features appearing in candidate sub-regions.
13. The system of claim 11 , comprising code for selecting sets of regionlets from a pool of candidate regionlet groups by boosting.
14. The system of claim 11 , comprising code for using a regionlet representation to model fine-grained spatial appearance layouts.
15. The system of claim 11 , comprising RealBoost code to train cascaded classifiers for object detection.