For image distortion classification, this paper proposed an algorithm based on Gabor wavelet and CNN. It used the good characteristic of Gabor wavelet to extract rough feature of images firstly, and then used the improved CNN to extract the key feature from rough feature. The main steps included pre-processing image firstly(including labels setting, samples balance and samples expansion); then it calculated eight directions Gabor wavelet to preprocessed images, and added eight sub-bands to one sample for training; finally, it used a self-designed CNN and Softmax classifier to train the final model, and used the methods of random gradient descent and error back propagation to optimize the parameters of convolution kernels during training. Using the final model to determine the type of image distortion, the classification accuracy on the LIVE standard image library is 95.62%. It shows that the proposed method has high accuracy and robustness.

[2019-10-05] Printed Article /doc/2018.03.0231/v2