A Neural Network for Denoising Fringe Patterns with Nonuniformly Illuminating Background Noise

There are many interference fringe patterns with nonuniformly illuminating background noise. Nonuniformly illuminating background noise may arise from diverse sources, such as scratches, small interfering concentric circles, gaussian noise, etc. with a nonuniform light source. In this work, we propose a practical fringe pattern denoising method using a simple neural network. However, nonuniform illumination prevents us from comparing the denoised images with the ground truths. The verification of the proposed method is shown indirectly using the uniformized results obtained from a nonuniform illumination correcting process.