P Gürdal1, C F Hildebolt, B G Akdeniz. 1. Department of Oral Diagnosis and Radiology, School of Dentistry, Ege University, Bornova, Izmir, Turkey.
Abstract
OBJECTIVES: To determine if variations in radiodensity data are introduced by lossy Joint Photographic Experts Group (JPEG) compression and/or the use of three software programs. METHODS: An occlusal film with an aluminium step wedge was exposed, processed and digitized under standard conditions. Before the image was saved, the coordinates and the gray-scale value for each pixel in a 20 x 20 pixel area near the middle of the thickest step were recorded. These pixel coordinates and gray-scale values served as Truth 1. In addition, a digital simulated-radiographic image with assigned gray-scale values for each pixel was created and served as Truth 2. The digital data for the scanned radiograph and the simulated radiograph were saved as Tagged Image File Format (TIFF) and lossy JPEG files. Each file was opened with three software programs and the gray-scale values of homologous pixels were recorded. For these pixels in each image type, the mean gray-scale values and standard deviations were calculated. The pixel gray-scale values for each homologous pixel were also individually compared. RESULTS: When the TIFF images were opened with the three software programs, one program resulted in gray-scale values that were not concordant with truth. All JPEG images resulted in gray-scale values that were not concordant with truth. CONCLUSIONS: One software program added a column of 0s to data files. Lossy JPEG compression introduced potentially deleterious variations to radiodensity data, and at least two of the software programs performed JPEG image decompression differently.
OBJECTIVES: To determine if variations in radiodensity data are introduced by lossy Joint Photographic Experts Group (JPEG) compression and/or the use of three software programs. METHODS: An occlusal film with an aluminium step wedge was exposed, processed and digitized under standard conditions. Before the image was saved, the coordinates and the gray-scale value for each pixel in a 20 x 20 pixel area near the middle of the thickest step were recorded. These pixel coordinates and gray-scale values served as Truth 1. In addition, a digital simulated-radiographic image with assigned gray-scale values for each pixel was created and served as Truth 2. The digital data for the scanned radiograph and the simulated radiograph were saved as Tagged Image File Format (TIFF) and lossy JPEG files. Each file was opened with three software programs and the gray-scale values of homologous pixels were recorded. For these pixels in each image type, the mean gray-scale values and standard deviations were calculated. The pixel gray-scale values for each homologous pixel were also individually compared. RESULTS: When the TIFF images were opened with the three software programs, one program resulted in gray-scale values that were not concordant with truth. All JPEG images resulted in gray-scale values that were not concordant with truth. CONCLUSIONS: One software program added a column of 0s to data files. Lossy JPEG compression introduced potentially deleterious variations to radiodensity data, and at least two of the software programs performed JPEG image decompression differently.
Authors: Mahitab M Soliman; Azza Abdulrahman Zaki; Hanaa Mohamed El Gazaerly; Ammar Al Shemmrani; Abd El Latif Sorour Journal: Int J Health Sci (Qassim) Date: 2014-10