A Comparative Analysis of FFT, DCT, and Wavelet-based Image Compression on Kaggle Rose Images
Salem, Nema Shalabi, Shefaa Khawfani, Ruba Salem, Nema
Salem, Nema
Shalabi, Shefaa
Khawfani, Ruba
Salem, Nema
Citations
Altmetric:
Type
Supervisor
Date
2025
Collections
Research Projects
Organizational Units
Journal Issue
Abstract
This chapter presents a comprehensive comparative analysis of three widely used transformbased image compression techniques: Fast Fourier Transform (FFT), Discrete Cosine Transform (DCT), and Wavelet Transform. The compression transforms are tested under three threshold settings. Utilizing a dataset of 497 JPEG images from Kaggle, we systematically evaluated each method under three distinct threshold settings to examine their performance. The ROSE images dataset was chosen due to its abundant edges and intricate details, effectively distinguishing the differences between compressed images. Setting A employed the highest threshold value, resulting in minimal compression. Setting B had an intermediate threshold value, while Setting C utilized the lowest threshold value, leading to maximum compression. Lower threshold values discard more coefficients, thereby increasing the compression ratio, but potentially affecting image quality. The evaluation metrics used in this study include Mean Squared Error (MSE), Peak signal-to-noise ratio (PSNR), Signal-to-quantization noise ratio (SQNR), Structural Similarity Index (SSIM), Normalized Cross-Correlation (NCC), Compression Ratio (CR), Memory saving (MS), and contrast. The inherent properties of the Wavelet Transform, which efficiently captures both frequency and spatial information, suggest it has a greater potential to excel in balancing compression efficiency and image quality preservation. Both FFT and DCT transforms are anticipated to deliver moderate performance in terms of the performance metrics. They may exhibit more variability and less effectiveness in retaining high-frequency image details. These insights underscore the importance of selecting the appropriate compression technique based on the specific requirements of image quality and storage efficiency, particularly highlighting the advantages of the Wavelet Transform for high-fidelity applications such as medical imaging and archival storage.
Department
Publisher
Sponsor
Copyright
Book title
Mathematical Modelling for Big Data Analytics