<p>Periodic patterns play a crucial role in imparting unique functionalities to surfaces, such as superhydrophobicity, anti-fouling properties, and color conversion. Traditional methods for aligning patterns often result in significant defects due to larger seam sizes and the need for tailored alignment techniques. To address these challenges, this study proposes a novel deep learning-based method for precisely aligning periodic patterns using image generation and Fourier Transform. The proposed method involves generating a large dataset of images with controlled rotational misalignments and applying Fourier Transform to convert these images into a frequency domain representation. Two strategies were employed: one utilizing overlapped images and the other using Fourier-transformed images. The Fourier-transformed images exhibited superior accuracy in predicting alignment errors, benefiting from their sensitivity to rotational changes. Several deep learning models, including custom convolutional neural networks (CNNs), ResNet50, and EfficientNet-b0, were trained and evaluated. The EfficientNet-b0 model achieved the highest accuracy with an RMSE of 0.018, significantly outperforming traditional methods. The model was tested on real-time video data, successfully predicting alignment angles within trained ranges, demonstrating its practical applicability in industrial settings. This deep learning approach eliminates the need for alignment marks and manual image analysis, offering a scalable solution for precisely aligning complex patterns in manufacturing processes.</p>