CNN-Based Fringe Alignment for Optical Experiments

Created using ChatSlide

This research focuses on a CNN-based approach for resolving misalignment issues in fringe patterns caused by optical imperfections. By introducing a two-stage CNN architecture, coarse corrections are first applied using ResNet for significant deviations, followed by fine alignment with a lightweight regression CNN. The theory underlying fringe formation and visibility serves as a foundation for this methodology. Data for model training comprises simulated fringe patterns with known angle...