This study presents a comparative analysis of the performance of two methods for calculating the sum in a convolution matrix 3 × 3 for image processing: The parallel method and the proposed dynamic method. The experiments covered a wide range of image sizes (from 100 × 100 to 500 × 500) and processor core counts (from 2 to 6). The results demonstrate that the dynamic method has significantly lower computational complexity (((×ℎ+×ℎ) × 3) compared to the parallel method ((×ℎ×2 × 8)) and achieves higher performance. The dynamic method remains efficient for both small and large images, with minimal dependence on the number of processor cores, making it a universal solution for scalable data processing. While the parallel method achieves considerable acceleration with an increasing number of cores, it exhibits limited scalability for larger images. The experimental findings highlight the advantages of the dynamic approach for image processing tasks where execution speed and resource efficiency are critical.