In the dilated convolution
of , the same filter can be applied at different receptive fields using different dilation factors by skipping a certain number of input values when applying the filter to a convolutional layer.
For each convolution
position on the image a set of 9 anchors having three aspect ratios, three scales and sharing a common center.
Such an interaction may be referred to as a convolution
of the original functions.
 designed a convolution
neural network named DrainNet for removing rain streaks from single image.
By increasing the convolution
block of MKNet-A, the classification accuracy can be gradually increased.
Each neuron in convolutional layer is only partially connected to input of current layer, which is equivalent to using a convolutional kernel to perform ergodic convolution
to the input feature figure of current layer.
The CNN contains unknown parameters: the filtering coefficients and biases in convolution
layers and the weights and biases in fully connected layers.
In the generative network, which is shown in the first row of Figure 3, there are two modules of convolution
Equation (2) is in fact a two-dimensional convolution
product that can also be assessed in the frequency domain.
Generally, the convolution
layer is connected to the input layer in a convolutional neural network (CNN).
YOLO does not use region proposal, but directly convolution
operations on the whole image, so it is faster than Faster-RCNN in speed, but the accuracy is less than Faster-RCNN.
Ruscheweyh and Sheil-Small  proved that the Hadamard product or convolution
of two analytic convex functions is also convex analytic and that the convolution
of an analytic convex function and an analytic close-to-convex function is close-to-convex analytic in the unit disk E.