Improving transfer learning accuracy by reusing Stacked Denoising Autoencoders | IEEE Conference Publication | IEEE Xplore