Image Definition Identification Algorithm Based on Lifting Wavelet Transform and Naive Bayes Classifier
Ling Qi, Yong Li Zhao, Ling Gao, Wei Wang · Applied Mechanics and Materials · 2013
Identification of definition for digital image is an important aspect of digital imaging system. To improve the efficiency of the present image definition identification methods with a high accuracy, an algorithm based on the compound model of Lifting Wavelet Transform and Naive Bayes classifier is proposed. Firstly, the two-dimensional Lifting Wavelet Transform is used to extract the image feature, and 28 statistical values obtained from 7 wavelet components by statistical process are treated as image eigenvalues for the follow-up identification. Then Naive Bayes classifier is used to achieve the identification, which has a high computational efficiency and competitive accuracy, and the classifier applied to the experiments of this paper is from OpenCV. The experiment consists of two phases. In phase one, the compound model is trained by 200 images from the training set. Similarly, in phase two, the model is tested by 100 images from the testing set. The results show that the algorithm based on the compound model is very effective, and obtains a high recognition rate.