Digital Radiography using High Dynamic Range Technique DAN CIURESCU 1, SORIN BARABAS 2, LIVIA SANGEORZAN 3, LIGIA NEICA 1 1 Department of Medicine, 2 Department of Materials Science, 3 Department of Computer Science Transilvania University of Brasov 25, Eroilor, blvd, 500030, Brasov ROMANIA ciurescu@yahoo.com, ab.sorin@gmail.com, sangeorzan@unitbv.ro, ligianeica@yahoo.com http://www.unitbv.ro Abstract: - This paper present benefits of High Dynamic Range (HDR) method. This technique improves the quality of digital We present the result of processing the input images using high dynamic range software. It is noticed, that in the final image (processed image), by applying the filter, on can distinguish details that in the input images cannot be clearly observed. Key-Words: - High Dynamic range (HDR), Digital Radiography, Filter, icam06, Photomatix, Essential HDR 1 Introduction This paper outlines the significant benefits of High Dynamic Range (HDR) proceeding in radiography reading and wants to show how existing HDR techniques improve visibility range in digital HDR technique encode the full visible range of luminance and color gamut, thus offering ultimate fidelity, limited only by the capabilities of the human eye and not by any existing technology. Also, enhance details situated in shadow and highlights of digital On can build a new image, using a HDR technique, by capture multiple photographs with different exposure times or by using high dynamic range sensors. Synthetic HDR images can be generated with computer graphics [3]. This technique is implemented in some photo software like Essential HDR, Photomatix Pro and others tools. 2 Theoretical Aspects High dynamic range (HDR) in image framework is a set of techniques that accurately represent the real scenes, emphasizing the shadow and highlight zones. This method consist in taking of two or more images of same object, with different parameters of exposure and then in composition into a final image or in capturing a single image from a sensor with native high dynamic range. HDR technique based on taking several images was developed by Paul Debevec [6] in 1997. Kuang, Johnson and Fairchild, director at Munsell Color Science Laboratory, proposed in 2006, new photo software using icam06 model, which incorporates the processing models in the human visual system that enhance local details. Photo soft-wares for processing the images are: - Photomatix Pro, this is a program that creates and processes HDR images; - Essential HDR is a HDR tone mapping software, developed by Imaging Luminary. 2.1 Digital image processing In digital photography, the range of exposure values which can be captured correctly by the sensor is also limited. The voltage generated by a CCD or CMOS is, to a good approximation, a linear function of the exposure value. CCD, Charge Coupled Device, is a video image sensor chip and is the most widely used sensor format. CMOS, Complementary Metal Oxide Semiconductor is also a video image chip, but with much lower quality picture than CCD chip. Unlike the overexposed pixels are replaced by a uniform white area with a total loss of detail. Although the sensor response is theoretically linear, for low exposure values, the quantum noise in the incident light, the thermal noise inside the sensor circuitry and finally the quantization noise introduced by the converter can surpass the signal intensity and compromise the visibility of the details ([1]). HDR image is made from several different images or a render from a render engine that supports unclamped luminance values, illustrated in Fig.1. For displaying HDR images is used a nonlinear bilateral filter. Relationship between luminance ([5]), contrast and colorfulness is described by Stevens and Hunt ([2]). ISSN: 1790-2769 599 ISBN: 978-960-474-124-3
Fig. 1 Luminance values for bitmap and HDR images Hunt effect Hunt effect described the effects of dark and light adaptation on color perception Fig. 3 Contrast varies function of luminance ([9]) Durand- and Dorsey- filter Durand- and Dorsey-filter (bilateral filter) is a nonlinear filter that smoothes a signal while preserving strong edges ([13]). This filter used a two-scale decomposition of the image into a base layer and a detail layer ( [7]). The method is fast and requires no parameter setting. 3 icam06 model in HDR rendering icam06 is a new image appearance model developed for the applications of HDR. This model was developed by Fairchild and Johnson ([9], [10], [11] and [12]). Figures 4 present a general flowchart of icam06 for HDR image rendering. Fig. 2 Saturation varies function of luminance ([9]) The image elements will appear more striking if the image.is moved to a significantly brighter viewing environment. Stevens effect Stevens s effect refers to an increase in brightness contrast with increasing luminance as presented in Fig.3. The tristimulus values (XYZ) are input data for icam06 model that are the stimulus image or scene in absolute luminance units ([12]). To predict various luminancedependent phenomena, such as the Hunt effect and the Stevens effect ( [8]), it is necessary to know the absolute luminance Y of the image data. The input image (represented in device independent coordinates) is decomposed into a base layer and detail layer. The base layer is obtained with an edge-preserving filter proposed by Durand and Dorsey ([4]). The detail layer is obtained by subtracting the base layer image from the original image ([4]). ISSN: 1790-2769 600 ISBN: 978-960-474-124-3
Example 1 A suffering tooth was X-rayed and its image is presented in two captures. One capture is underexposed (fig.5.a) and the other, respectively, overexposed (fig5.b). Fig.5 Input images, underexposed (a) Fig.6. Radiological normal image Fig. 4 Model of icam06 for HDR image rendering 4 Case study Using Photomatix Pro- and Essential HDR-software, we present two types of radiography, namely: At the same time we have a normal image of tooth (Fig.6) that we can compare with images resulted after HDR processing (Fig.7.a with Photomatix and Fig. 7.b with Essential HDR photo software). - Dental radiography - Medical radiography The processed images were made with digital radiological equipment. Dental radiography For a suffering tooth on made two images. On wish to detect the affected region. The images are processed using the two photo softwares, Photomatix and Essential HDR. On present two examples for different tooth dental radiography processed with the above named software. Fig.7.Output images, with Photomatix (a) and with Essential HDR (b) ISSN: 1790-2769 601 ISBN: 978-960-474-124-3
Example 2 Another dental radiography is presented in two captures: underexposed (Fig.8.a) and overexposed (Fig8.b). Also, the difference between the two processing (with Photomatix and Essential HDR) is due to a tonal adjustment and detail adjustment. Medical radiography In studying the bones of the pelvis on made two images. On wish to detect the affected region. The images are processed using the Essential HDR. Exemple3 Fig. 8 Input images, underexposed (a) Fig.11 Input images, underexposed (a) Fig.9 Radiological normal image Fig.10.Output images, with Photomatix (a) and with Essential HDR (b) At the same time we have a normal image of tooth (Fig.9) that can be compared with images resulted after HDR processing (Fig.10). It is noticed that in the final image (processed image) by applying the filter it can be noticed details that in the other tree images cannot be distinguished Fig.12 Normal image (a) and image with Photomatix (b) The resulting image (Fig.12.b), using Photomatix software, is more clear and with more details that the other images (Fig.11, Fig.12.a). 5 Conclusion High dynamic range images (HDR) are images that have a large contrast ratio. Using Durand filter, tone mapping is necessary to display these images on a device with a much smaller dynamic range. ISSN: 1790-2769 602 ISBN: 978-960-474-124-3
Experimental results have further demonstrated that the new methods can produce good results on a variety of high dynamic range images. Both HDR software are very good and are coming to the help of the physician, to interpret correctly the The new icam06 model has been tested on a variety of radiography using two softwares namely, Essential HDR and Photomatix Pro. The HDR digital radiographic imaging has in dental and medical practice a great importance. References: [1] Peter G. J. Barten. Contrast Sensitivity of the Human Eye and its Effects on Image Quality. SPIE Optical Engineering Press, 1999. [2] Gunter Wyszecki and W. S. Stiles. Color Science: Concepts and Methods, Quantitative Data and Formulae (2nd edition). John Wiley & Sons, 1982. [3] Livia Sângeorzan, Parpalea Mircea. Some Aspects in the Modelling of Physic Phenomena using Computer Graphic, 2008 Proceedings of the 10 th WSEAS International Conference on Mathematical and Computational Methods in Science and Engineering (MACMESE'08), pp. 518-523. [4] Gabriele Guarnieri, Luigi Albani, Giovanni Ramponi, Minimum-Error Splitting Algorithm for a Dual Layer LCD, Journal of Display Technology, Vol. 4, Issue 4, pp. 383-390 [5] Gabriele Guarnieri, Giovanni Ramponi, Silvio Bonfiglio, and Luigi Albani. Nonlinear mapping of the luminance in dual-layer high dynamic range displays. Proc. Electronic Imaging 2009, Image Processing: Algorithms and Systems VII, volume 7245, page 72450D. SPIE, 2009 [6] Paul E. Debevec and Jitendra Malik., Recovering high dynamic range radiance maps from photographs. SIGGRAPH 97, August 1997. [7] F.Durand and J. Dorsey, Fast Bilateral Filtering for the Display of High-Dynamic Range Images, Proceedings of SIGGRAPH 02, 257-266 (2002). [8] M.D. Fairchild and G.M. Johnson, Meet icam: An image color appearance model, IS&T/SID 10th Color Imaging Conference, 33-38 (2002). [9] Mark D. Fairchild, Color Appearance Models, second edition 2004 pp121-123 [10]. Fairchild, M.D. and Johnson G.M., 2004. The icam framework for image appearance, image differences, and image quality, J. of Electronic Imaging 13,pg. 126-138. [11]. Johnson, G.M. AND Fairchild, M.D. 2003. Rendering HDR images. IS&T/SID 11th Color Imaging Conference, Scottsdale, pg. 36-41. [12]. Jiangtao Kuang, Mark D. Fairchild icam06, HDR, and Image Appearance, Rochester Institute of Technology, Rochester, New York. [13] Sylvain Paris, Fredo Durand, A Fast Approximation of the Bilateral Filter Using a Signal Processing Approach, International Journal of Computer Vision, Vol.18, Issue 1 (January 2009), pg.24-52, Kluwer Academic Publisher Hingham, MA, USA ISSN: 1790-2769 603 ISBN: 978-960-474-124-3