JPEG2000 2D and 3D reversible compressions of thin-section chest CT images: improving compressibility by increasing...

To propose a preprocessing technique that increases the compressibility in reversible compressions of thin-section chest computed tomographic (CT) images and to measure the increase in compression ratio (CR) in Joint Photographic Experts Group (JPEG) 2000 two-dimensional (2D) and three-dimensional (3D) compressions.

This study had institutional review board approval, with waiver of informed patient consent. A preprocessing technique that automatically segments pixels outside the body region and replaces their values with a constant value to maximize data redundancy was developed. One hundred CT studies (50 standard-radiation dose and 50 low-radiation dose studies) were preprocessed by using the technique and then reversibly compressed by using the JPEG2000 2D and 3D compression methods. The CRs (defined as the original data size divided by the compressed data size) with and those without use of the preprocessing technique were compared by using paired t tests. The percentage increase in the CR was measured.

The CR increased significantly (without vs with preprocessing) in JPEG2000 2D (mean CR, 2.40 vs 3.80) and 3D (mean CR, 2.61 vs 3.99) compressions for the standard-dose studies and in JPEG2000 2D (mean CR, 2.38 vs 3.36) and 3D (mean CR, 2.54 vs 3.55) compressions for the low-dose studies (P < .001 for all). The mean percentage increases in CR with preprocessing were 58.2% (95% confidence interval [CI]: 53.1%, 63.4%) and 52.4% (95% CI: 47.5%, 57.2%) in JPEG2000 2D and 3D compressions, respectively, for the standard-dose studies and 41.1% (95% CI: 38.8%, 43.4%) and 39.4% (95% CI: 37.4%, 41.7%) in JPEG2000 2D and 3D compressions, respectively, for the low-dose studies.

The described preprocessing technique considerably increases CRs for reversible compressions of thin-section chest CT studies.