The Comparison of 2D Convolution and Max Pooling Process in Real Time

Convolution and max pooling process are the necessary processes to get an object’s feature in convolutional neural network (CNN). There are many filters or kernels to be convolved with the input image, to get another form of the object’s edge. The max pooling is one way to reduce the spatial dimension. In this paper, we study a comparison of 2D convolution and max pooling process in real time. The comparison are the process to get the output image. The first one, we perform convolution first, then the max pooling process. The second one is performing the max pooling first, then convolution process. Both process is required to get the object’s feature to be fed to classifier. The experiment is performed using programming language C++ and openCV library.