Gaussian Blur Optimization Method: Linear Texture Filtering Under Gaussian Function Separation
Shangdong Yang, Jue Wu · 2024
Gaussian blur technology has been widely used in computer graphics and vision. Although the traditional Gaussian blur method can achieve the blurring effect of the image, it has high computational cost and slow rendering efficiency when processing high-resolution images or continuously rendering multiple images. This paper proposes a method based on the separability and semigroup properties of the Gaussian function, combined with linear texture filtering, to extract and process texture information, which can effectively speed up the sampling speed, reduce the cost of convolution operations, and improve rendering efficiency. Experimental results show that under the WebGL (Web Graphics Library) 3D rendering protocol technology, the efficiency of this algorithm is particularly significantly improved in the continuous rendering of high-resolution images, showing good application prospects.