Watermarking Representation for Adaptive Image Classification with Radial Basis Function Network
Chi‐Man Pun · Machine Learning · 2010
In this chapter, a novel approach using watermarking representation for adaptive image classification with Radial Basis Function (RBF) network has been proposed. Experimental results show that the proposed method has strong resistance to noise and JPEG compression with very low quality factor, and has much better efficiency than the other image classification method. However, the performance for median filtering attacks still needs to be improved further. For image classification experiments, the best performance was obtained using only 47 features with 96.8% accuracy. Future work may focus on embedding more useful image features such as invariant features for image analysis.