A comparative study of Singular Value Decomposition (SVD) and Discrete Cosine Transform (DCT) techniques for image compression applications
Palavras-chave:
Image Compression, Singular Value Decomposition (SVD), Discrete Cosine Transform (DCT)Resumo
The demand for methods capable of extracting knowledge from data and organizing large amounts of
information has been growing in recent decades. Images are present in several applications in biomechanics and
biomedicine, and usually require the use of computational methods to reduce complexity without compromising
quality. This paper proposes the application of computational methods for image compression, considering data
storage and transfer issues. In particular, we present a comparative study between the Singular Value Decompo-
sition (SVD) and Discrete Cosine Transform (DCT) techniques for image compression. We use a referencial way
of obtaining some metrics for compression without excessive loss of information. The computational experiments
in this work consisted of fixing compression metrics and evaluating the performances of DCT and SVD. The re-
sults demonstrate that DCT presents, in most tests, better performance in reducing memory usage in compression
compared to SVD, despite resulting in lower image quality.
