Fourier Transform-Based Method for Removing Image Textures

This article presents a Fourier Transform-based method for removing image textures. Traditional image processing techniques often leave residual highlight regions when eliminating textures, compromising image clarity and quality. The proposed method converts images to the frequency domain using Fast Fourier Transform (FFT), where high-frequency components corresponding to texture patterns are identified as bright spots in the spectral magnitude. By applying a frequency domain filter (such as a Gaussian low-pass filter or notch filter) to suppress these high-frequency regions, the method effectively removes textures while preserving low-frequency structural information. Key implementation steps include: 1) Converting the image to grayscale and applying 2D FFT, 2) Creating a frequency mask to attenuate high-frequency components, 3) Applying inverse FFT to reconstruct the texture-reduced image. The algorithm demonstrates high computational efficiency due to optimized FFT operations and produces superior results suitable for practical applications in medical imaging, document processing, and computer vision systems.