Wavelet transform-based multi-column convolutional neural network copied picture detection method

The invention provides a wavelet transform-based multi-column convolutional neural network copied picture detection method. The method comprises the steps: acquiring a to-be-detected image and preprocessing the to-be-detected image; performing Haar wavelet transform on the preprocessed to-be-detected image to obtain low-frequency data and high-frequency data; inputting the low-frequency data and the high-frequency data into a pre-trained multi-column convolutional neural network to obtain a density map; and inputting the density map into a pre-trained convolutional neural network to obtain a first probability that the to-be-detected image is a copied image, and determining whether the to-be-detected image is the copied image according to the first probability. According to the method, the high-frequency features of the to-be-detected image are effectively extracted through Haar wavelet transform, and the generalization ability of the convolutional neural network is improved; and moreover, the multi-column con