EscherNet 101
This addresses a domain-specific problem in pattern recognition and symmetry analysis, with incremental contributions in filter analysis.
The paper tackled the problem of classifying 2D periodic patterns into 17 wallpaper groups using a deep learning model called EscherNet 101, achieving results measured by classification rates and analyzing learned filters to capture second-order invariants.
A deep learning model, EscherNet 101, is constructed to categorize images of 2D periodic patterns into their respective 17 wallpaper groups. Beyond evaluating EscherNet 101 performance by classification rates, at a micro-level we investigate the filters learned at different layers in the network, capable of capturing second-order invariants beyond edge and curvature.