3. Introduction
Detection의 종류
2019-07-15
3
Detection
with DL
1 Stage
method
2 Stage
method
• 진정한 의미의 end-to-end 구조
• 빠른 detection 속도
• eg. YOLO 계열
• Region proposal and detection 구조
• 속도는 느리나 정확도가 높음
• eg. R-CNN 계열
R-FCN(Region-based Fully Convolutional Networks)은 R-CNN 계열에 FCN 구조를 접목함.
12. R-FCN
Position-Sensitive RoI Pooling
2019-07-15
12
+
+
=
)
,
(
)
,
(
0
0
,
, /
)
|
,
(
)
|
,
(
j
i
bin
y
x
c
j
i
c n
y
y
x
x
z
j
i
r
0, 0 0, 1 0, 2
1, 0 1, 1 1, 2
2, 0 2, 1 2, 2
bin(i,j)
)
|
,
(
j
i
rc
)
|
0
,
0
(
person
r )
|
2
,
2
(
person
r
)
|
0
,
0
(
car
r )
|
2
,
2
(
car
r
…
…
…
c
j
i
z ,
, : score map
]
)
1
[(
]
[
k
w
i
x
k
w
i +
]
)
1
[(
]
[
k
h
j
y
k
h
j +
13. R-FCN
Position-Sensitive RoI Pooling
2019-07-15
13
]
)
1
[(
]
[
k
w
i
x
k
w
i +
]
)
1
[(
]
[
k
h
j
y
k
h
j +
w
h
0 1 2
0
1
2
i
j
=
j
i
c
c j
i
r
r
,
)
|
,
(
)
(
=
= C
c
r
r
c
c
c
e
e
s
0
'
)
(
)
(
'
)
(
+
+
=
)
,
(
)
,
(
0
0
,
, /
)
|
,
(
)
|
,
(
j
i
bin
y
x
c
j
i
c n
y
y
x
x
z
j
i
r
14. R-FCN
Training
2019-07-15
14
)
,
(
]
0
[
)
(
)
,
( *
*
,
,
, * t
t
L
c
s
L
t
s
L reg
c
cls
h
w
y
x
+
=
*
c : RoI’s ground-truth label
)
log(
)
( *
*
c
c
cls s
s
L −
= : Cross entropy loss for classification
*
t : Ground-truth box
reg
L : bounding box regression loss
>0.5 : positive
o.t : negative
GT
Predict
Softmax
score map
Regression
loss(IoU)