Accelerating Image Preprocessing with CUDA: High-Speed Gaussian Filtering and Brightness Enhancement

These days, we can see that as technology advances, so does the volume of data. Furthermore, it takes a long time to perform digital operations on image-based data. This article examines issues related to computing devices' speed at processing massive volumes of data. Using one of the image preprocessing techniques, the Gaussian filter, the study's primary objective is to speed up the processes of enhancing image brightness and eliminating noise. This is accomplished by processing images of various sizes in parallel on the computer's graphics processor using CUDA (Compute Unified Device Architecture) technology, as well as on the central processor using the OpenMP (Open Multi-Processing) parallel library for devices without a graphics processor. The study concludes with the presentation of the percentage increase indicators of CUDA technology in comparison to OpenMP when preprocessing images of sizes 512x512, 1024x1024, and 3200x2400 using these parallel processing technologies. At the end of the study, the capabilities of CUDA technology in image processing are highly evaluated compared to other parallel processing tools. OpenMP technology is proposed for parallel image processing on devices without a graphics processor.