A signal processing interpretation of noise-reduction convolutional neural networks
This work provides a theoretical foundation for designing noise-reduction CNNs, which is incremental as it unifies existing scattered ideas to make them more accessible.
The paper tackles the lack of theoretical underpinnings for noise-reduction convolutional neural networks (CNNs) by building intuition on deep convolutional framelets to explain diverse encoding-decoding CNN architectures in a unified framework, connecting signal processing principles to deep learning to offer guidance for designing robust and efficient novel CNNs.
Encoding-decoding CNNs play a central role in data-driven noise reduction and can be found within numerous deep-learning algorithms. However, the development of these CNN architectures is often done in ad-hoc fashion and theoretical underpinnings for important design choices is generally lacking. Up to this moment there are different existing relevant works that strive to explain the internal operation of these CNNs. Still, these ideas are either scattered and/or may require significant expertise to be accessible for a bigger audience. In order to open up this exciting field, this article builds intuition on the theory of deep convolutional framelets and explains diverse ED CNN architectures in a unified theoretical framework. By connecting basic principles from signal processing to the field of deep learning, this self-contained material offers significant guidance for designing robust and efficient novel CNN architectures.