Abstract
Appropriate selection of features may lead to the specificity of classification methods and identify the most critical features from all sparse or dense impact data using a filter based on the recognition selection method characterized. Filtration is used to reduce sample complexity, improve the clarity of viscous samples, and reduce background signals, resulting in increased signal-to-noise ratios in analytical tests. Depending on the filtration method applied, particles are separated based on properties such as size. This study assessed the impact of filter selection and the variation in the number of projections on the final reconstructed artificial phantom images. Utilizing image reconstruction techniques, it delves into the application of mathematical transforms, including Radon and Fourier, to improve image quality and resolution, particularly in medical imaging modalities such as CT and MRI. The research predominantly focuses on the application of the Filtered Back Projection (FBP) algorithm to reconstruct images from changing numbers of projections. The results underscore the main role of filter choice in removing noise, with the Ramp filter presenting the most promising results. The investigation concludes that reducing the number of projections results in a decline in image contrast and an increase in image noise.
