Net2Vis -- A Visual Grammar for Automatically Generating Publication-Tailored CNN Architecture Visualizations
This addresses the need for researchers in computer vision and deep learning to efficiently produce standardized and error-free CNN visualizations for publications, though it is incremental as it automates an existing manual process rather than introducing a new paradigm.
The authors tackled the problem of manually creating inconsistent and time-consuming visualizations of CNN architectures for publications by developing Net2Vis, an automated tool that translates Keras network specifications into publication-ready figures using a visual grammar derived from analyzing ICCV and CVPR papers from 2013 to 2019, resulting in reduced time and unified designs as validated through expert feedback and quantitative study.
To convey neural network architectures in publications, appropriate visualizations are of great importance. While most current deep learning papers contain such visualizations, these are usually handcrafted just before publication, which results in a lack of a common visual grammar, significant time investment, errors, and ambiguities. Current automatic network visualization tools focus on debugging the network itself and are not ideal for generating publication visualizations. Therefore, we present an approach to automate this process by translating network architectures specified in Keras into visualizations that can directly be embedded into any publication. To do so, we propose a visual grammar for convolutional neural networks (CNNs), which has been derived from an analysis of such figures extracted from all ICCV and CVPR papers published between 2013 and 2019. The proposed grammar incorporates visual encoding, network layout, layer aggregation, and legend generation. We have further realized our approach in an online system available to the community, which we have evaluated through expert feedback, and a quantitative study. It not only reduces the time needed to generate network visualizations for publications, but also enables a unified and unambiguous visualization design.