Characterization of CT Cancer Lung Image Using Image Compression Algorithms and Feature Extraction

Abstract

Image compression techniques find an extensive role in the field of biomedical image processing. Transform based image compression algorithms efficiency is mainly depending on the decoding methods, adopted. In this work, wavelet Transform based compression algorithms are developed for computer tomography image. Symlet based transformation of the CT images of lung are proposed in this work for the decomposition of the CT medical image. The decomposed images are encoded using the various encoding techniques such as Embedded Zero wavelet, (EZW), Set Partitioning in Hierarchical Trees (SPIHT).The developed compression algorithms are evaluated in terms of PSNR, Compression ratio, Means square error and bits per pixel. The optimum compression algorithm is also found based on the results obtained, so as to characterize the CT image the features are extracted and it is proven that after compression, the CT images show its ability for identifying types of defects. The results are an indicator to the promising application of this for medical image compression schemes. This paper provides the approach and analysis methodologies and the results obtained.