Information set supported deep learning architectures for improving noisy image classification

It extends the classic MNIST dataset by incorporating a more diverse set of pattern classes. This shows that the information layer introduced in standard CNN helps to boost the high level features. Figure 4 illustrates the confusion matrices for MNIST dataset comparing CNN and ISCNN-III performance with varying PSNR. When images are not contaminated by noise, the performance of classical CNN and information set theory-based CNN is almost similar. In Table 8, for the uncorrupted data (PSNR = Inf) the performance of both the architectures for MNIST dataset is identical.