Convolution Exploration Widget
image/svg+xml
A 3x3 convolutional filter ([[1, 0, 0], [1, 1, 0], [0, 0, 1]) being applied to the top-left of a 5x5 feature map ([3, 5, 2, 8, 1], [9, 7, 5, 4, 3], [2, 0, 6, 1, 6], [6, 3, 7, 9, 2], [1, 4, 9, 5, 1]), generating an output value of 25 in the top-left of the 3x3 output feature map.
Input Feature Map
Output Feature Map
8
1
4
3
1
6
6
3
7
9
2
1
4
9
5
1
3+0+0+9+7+0+0+0+6
3×1
5×0
2×0
9×1
7×1
5×0
2×0
0×0
6×1
18
17
18
22
14
20
15
23
25
image/svg+xml
A 3x3 convolutional filter ([[1, 0, 0], [1, 1, 0], [0, 0, 1]) being applied to the top-middle of a 5x5 feature map ([3, 5, 2, 8, 1], [9, 7, 5, 4, 3], [2, 0, 6, 1, 6], [6, 3, 7, 9, 2], [1, 4, 9, 5, 1]), generating an output value of 18 in the top-middle of the 3x3 output feature map.
Input Feature Map
Output Feature Map
1
3
6
6
3
7
9
2
1
4
9
5
1
5+0+0+7+5+0+0+0+1
5×1
2×0
8×0
7×1
5×1
4×0
0×0
6×0
1×1
17
18
22
14
20
15
23
25
18
3
9
2
image/svg+xml
A 3x3 convolutional filter ([[1, 0, 0], [1, 1, 0], [0, 0, 1]) being applied to the top-right of a 5x5 feature map ([3, 5, 2, 8, 1], [9, 7, 5, 4, 3], [2, 0, 6, 1, 6], [6, 3, 7, 9, 2], [1, 4, 9, 5, 1]), generating an output value of 17 in the top-right of the 3x3 output feature map.
Input Feature Map
Output Feature Map
6
3
7
9
2
1
4
9
5
1
2+0+0+5+4+0+0+0+6
2×1
8×0
1×0
5×1
4×1
3×0
6×0
1×0
6×1
18
22
14
20
15
23
25
18
17
3
9
2
5
7
0
image/svg+xml
A 3x3 convolutional filter ([[1, 0, 0], [1, 1, 0], [0, 0, 1]) being applied to the middle-left of a 5x5 feature map ([3, 5, 2, 8, 1], [9, 7, 5, 4, 3], [2, 0, 6, 1, 6], [6, 3, 7, 9, 2], [1, 4, 9, 5, 1]), generating an output value of 18 in the middle-left of the 3x3 output feature map.
Input Feature Map
Output Feature Map
9
2
1
4
9
5
1
9+0+0+2+0+0+0+0+7
9×1
7×0
5×0
2×1
0×1
6×0
6×0
3×0
7×1
22
14
20
15
23
25
18
17
18
3
5
2
8
1
4
3
1
6
image/svg+xml
A 3x3 convolutional filter ([[1, 0, 0], [1, 1, 0], [0, 0, 1]) being applied to the center of a 5x5 feature map ([3, 5, 2, 8, 1], [9, 7, 5, 4, 3], [2, 0, 6, 1, 6], [6, 3, 7, 9, 2], [1, 4, 9, 5, 1]), generating an output value of 22 in the center of the 3x3 output feature map.
Input Feature Map
Output Feature Map
2
1
4
9
5
1
7+0+0+0+6+0+0+0+9
7×1
5×0
4×0
0×1
6×1
1×0
3×0
7×0
9×1
22
14
20
15
23
25
18
17
18
3
5
2
8
1
3
6
6
9
2
image/svg+xml
A 3x3 convolutional filter ([[1, 0, 0], [1, 1, 0], [0, 0, 1]) being applied to the middle-right of a 5x5 feature map ([3, 5, 2, 8, 1], [9, 7, 5, 4, 3], [2, 0, 6, 1, 6], [6, 3, 7, 9, 2], [1, 4, 9, 5, 1]), generating an output value of 14 in the middle-right of the 3x3 output feature map.
Input Feature Map
Output Feature Map
1
4
9
5
1
5+0+0+6+1+0+0+0+2
5×1
4×0
3×0
6×1
1×1
6×0
7×0
9×0
2×1
22
20
15
23
25
18
17
18
14
3
5
2
8
1
6
9
2
3
7
0
image/svg+xml
A 3x3 convolutional filter ([[1, 0, 0], [1, 1, 0], [0, 0, 1]) being applied to the bottom-left of a 5x5 feature map ([3, 5, 2, 8, 1], [9, 7, 5, 4, 3], [2, 0, 6, 1, 6], [6, 3, 7, 9, 2], [1, 4, 9, 5, 1]), generating an output value of 20 in the bottom-left of the 3x3 output feature map.
Input Feature Map
Output Feature Map
5
1
2+0+0+6+3+0+0+0+9
2×1
0×0
6×0
6×1
3×1
7×0
1×0
4×0
9×1
22
15
23
25
18
17
18
14
20
3
5
2
8
1
9
7
5
4
3
1
6
9
2
image/svg+xml
A 3x3 convolutional filter ([[1, 0, 0], [1, 1, 0], [0, 0, 1]) being applied to the bottom-middle of a 5x5 feature map ([3, 5, 2, 8, 1], [9, 7, 5, 4, 3], [2, 0, 6, 1, 6], [6, 3, 7, 9, 2], [1, 4, 9, 5, 1]), generating an output value of 15 in the bottom-middle of the 3x3 output feature map.
Input Feature Map
Output Feature Map
1
1
0+0+0+3+7+0+0+0+5
0×1
6×0
1×0
3×1
7×1
9×0
4×0
9×0
5×1
22
23
25
18
17
18
14
20
15
3
5
2
8
1
9
7
5
4
3
2
6
6
2
image/svg+xml
A 3x3 convolutional filter ([[1, 0, 0], [1, 1, 0], [0, 0, 1]) being applied to the bottom-right of a 5x5 feature map ([3, 5, 2, 8, 1], [9, 7, 5, 4, 3], [2, 0, 6, 1, 6], [6, 3, 7, 9, 2], [1, 4, 9, 5, 1]), generating an output value of 23 in the bottom-right of the 3x3 output feature map.
Input Feature Map
Output Feature Map
1
6+0+0+7+9+0+0+0+1
6×1
1×0
6×0
7×1
9×1
2×0
9×0
5×0
1×1
22
25
18
17
18
14
20
15
23
3
5
2
8
1
9
7
5
4
3
2
6
4
0
3