Determination of Population Density Using Convoluted Neural Networks
In one embodiment, a method includes receiving an image on a computing device. The computing device may further execute a classification algorithm to determine whether a target feature is present in the received image. As an example, the classification algorithm may determine whether a building is depicted in the received image. In response to determining that a target feature is present, the method further includes using a segmentation algorithm to segment the received image for the target feature. Based on a determined footprint size of the target feature, a distribution of statistical information over the target feature in the image can be calculated.
1 . A method comprising:
receiving an image;
executing a classification algorithm to determine whether a target feature is present in the received image;
in response to determining that a target feature is present, using a segmentation algorithm to segment the received image for the target feature and determine a footprint size of the target feature; and
calculating a distribution of statistical information over the target feature based on the determined footprint size of the target feature.
2 . The method of claim 1 , wherein calculating the distribution of statistical information is based at least in part on a property that scales with the size of the target feature within an image.
3 . The method of claim 1 , wherein the image received is a satellite photo of a geographic region.
4 . The method of claim 1 , wherein the target feature is the presence of buildings in the image.
5 . The method of claim 4 , wherein the footprint size of the target feature is the total area of the buildings in the image.
6 . The method of claim 1 , further comprising using a convoluted neural network to remove noise or haze using an adaptive learnable transformer, wherein the adaptive learnable transformer is trained to remove noise or haze from pixels corresponding to the target feature.
7 . The method of claim 6 , wherein the adaptive learnable transformer is trained using image-level labeled data and a weakly-supervised classification algorithm.
8 . The method of claim 1 , wherein the classification algorithm is a weakly-supervised classification algorithm trained using image-level labeled data, without pixel-level labeled data.
9 . The method of claim 1 , wherein determining whether a target feature is present in the received image comprises:
for each pixel in the received image, determining, using a weakly-supervised classification algorithm, a per-pixel probability that the pixel corresponds to the target feature; and
determining an average of the per-pixel probabilities for the pixels in the received image.
10 . The method of claim 1 , wherein the classification algorithm is a weakly-supervised classification algorithm comprising a feedback loop to suppress irrelevant neuron activations of a convoluted neural network.
11 . The method of claim 1 , wherein the segmentation algorithm is a weakly-supervised segmentation algorithm trained using image-level labeled data, without pixel-level labeled data.
12 . The method of claim 1 , wherein the segmentation algorithm is a weakly-supervised segmentation algorithm trained to minimize a loss function
f
w
=
arg
min
∑
i
1
2
y
i
-
f
w
(
x
i
)
2
+
λ
w
2
wherein ƒ w is the transformation from input x to output ŷ parametered with w.
13 . The method of claim 1 , wherein the segmentation algorithm comprises a convoluted neural network with a plurality of layers.
14 . The method of claim 13 , wherein each layer comprises a plurality of neurons.
15 . The method of claim 14 , wherein for a particular layer 1 with input x 1 and target output y 1 , the convoluted neural network optimizes a target function
min ½∥ y l −ƒ w ( x l )∥ 2 +γ∥x l ∥ 1 , wherein:
ƒ w is the transformation from input x to output ŷ parametered with w; and
for a particular neuron x i l :
if
∂
y
l
-
f
w
(
x
l
)
2
∂
x
i
l
>
γ
then x i l is positively activated;
if
∂
y
l
-
f
w
(
x
l
)
2
∂
x
i
l
<
-
γ
then x i l is negatively activated;
else, neuron x i l is deactivated.
16 . The method of claim 13 , wherein the convoluted neural network comprises one or more stacks, wherein each stack comprises a feedback unit comprising one or more layers.
17 . The method of claim 16 , wherein each feedback unit allows neurons with positive gradients to be activated.
18 . The method of claim 16 , wherein the one or more layers comprise:
a feedback layer;
a rectified linear unit layer;
a batch normalization layer; and
a convolution layer.
19 . One or more computer-readable non-transitory storage media embodying software that is operable when executed to:
receive an image;
execute a classification algorithm to determine whether a target feature is present in the received image;
in response to determining that a target feature is present, use a segmentation algorithm to segment the received image for the target feature and determine a footprint size of the target feature; and
calculate a distribution of statistical information over the target feature based on the determined footprint size of the target feature.
20 . A system comprising: one or more processors; and a memory coupled to the processors comprising instructions executable by the processors, the processors being operable when executing the instructions to:
receive an image;
execute a classification algorithm to determine whether a target feature is present in the received image;
in response to determining that a target feature is present, use a segmentation algorithm to segment the received image for the target feature and determine a footprint size of the target feature; and
calculate a distribution of statistical information over the target feature based on the determined footprint size of the target feature.