Crowd counting is an important part of crowd analysis, which is of great significance to crowd control and management.The convolutional neural network (CNN) based crowd counting method is widely used to solve the problem ofinsufficient counting accuracy due to heavy occlusion, background clutters, head scale and perspective changes in crowdscenes. The multi-column convolutional neural network (MCNN) is a CNN-based method for crowd counting, whichadapts to head scale variation of crowd scenes by constructing multi-column convolutional neural network composing ofthree single-column networks corresponding to the convolution kernel with different sizes (large, medium and small).However, as the MCNN network is relatively shallow, its receptive field is also limited, which affects the adaptability tolarge scale variations. In addition, due to insufficient training data, it is necessary to carry out a pre-training strategieswhich pre-trains the single-column convolutional neural network individually and combines the cumbersome. In thispaper, a crowd counting method based on multi-column dilated convolutional neural network was proposed. Dilatedconvolution was used to enhance the receptive field of the network, so as to be better adaptive to the head scalevariations. The image patches were obtained by randomly clipping from the original training data set images in theprocess of each iterative training to further expand the training data, while the training could be achieved without tediouspre-training. The experimental results on ShanghaiTech public dataset showed that the accuracy of crowd countingproposed in this paper was better than that of MCNN, which proved that this method is more robust to head scalevariations in crowd scenes.
展开▼
