Ge Wang, PhD, Director SBES Division & ICTAS Center for Biomedical Imaging VT-WFU School of Biomedical Engineering & Sciences Virginia Tech, Blacksburg, VA,
Download ReportTranscript Ge Wang, PhD, Director SBES Division & ICTAS Center for Biomedical Imaging VT-WFU School of Biomedical Engineering & Sciences Virginia Tech, Blacksburg, VA,
Ge Wang, PhD, Director SBES Division & ICTAS Center for Biomedical Imaging VT-WFU School of Biomedical Engineering & Sciences Virginia Tech, Blacksburg, VA, USA [email protected] October 1, 2010 Source Interior Tomography (2007) Beam Beam ROI Object Object Regular Reconstruction Known Sub-region Beam Interior Problem Sparsity Model ROI Object Landmark-based Interior Tomography Beam ROI Object Sparsity-based Interior Tomography First Paper (May 2007) Independent Work (Oct. 2007) Interior CT Patent Literature Analysis Outline • Less Is Deeper • Less Is Larger • Less Is Faster • Less Is Less • Less Is More Less Is Deeper Use of less projection data for accurate image reconstruction demands deeper insight, more advanced theory and more powerful tools. Computed Tomography (Wholesale) Measurement y Sinogram t P( , t ) p x Object X-rays f ( x, y) Reconstruction t Inner Vision with Local Data (Retail) Measurement y Sinogram t P( , t ) x Object X-rays f ( x, y) X X Reconstruction t Earliest BPF Formula (1991) Half-PI-Line Reconstruction (2006) Field of View (FOV) Partial-PI-Line Extrapolation from a Known Point (2006) FOV ? Curved Filtering Path (2003) ? Interior Reconstruction Global FBP Local FBP Local SART Interior Recon (a) (b) (HU) (HU) 500 500 0 0 -500 -500 (c) -1000 50 100 150 200 (Pixel) (d) -1000 50 100 150 200 (Pixel) HOT Sparsity-based Interior Recon Outline • Less Is Deeper • Less Is Larger • Less Is Faster • Less Is Less • Less Is More Less Is Larger Acquisition of less projection data is achieved with a narrower beam, and an object larger than the beam width is not a concern. X-ray Beam Rotation Axis Preclinical Nano-CT Central Stop Sample Zone Plate Condense Lens ROI Sample Stage Phase Ring Potential for Study on Earliest Life Hagadorn JW, et al. (2006) Cellular and subcellular structure of Neoproterozoic embryos. Science 314:291–294 Big Patient Problem Outline • Less Is Deeper • Less Is Larger • Less Is Faster • Less Is Less • Less Is More Less Is Faster Less data means smaller detector size, faster frame rate, and more imaging chains, all of which contribute to accelerate the data acquisition process. Spiral Cone-beam CT Dual-source Clinical CT (2005) Multi-source Interior Tomography ROI Wang G, Yu H, Ye YB. Virginia Tech Patent Disclosure on May 15, 2007, US Patent Application 12/362,979 allowed on October 21, 2009 Ye YB, Yu HY, Wei YC, Wang G. International Journal of Biomedical Imaging, Article ID:63634, 2007 Wang G, Yu H, Ye Y. Medical Physics. 36:3575-3581, 2009 From Scanning to Roaming Outline • Less Is Deeper • Less Is Larger • Less Is Faster • Less Is Less • Less Is More Less Is Less Less data is equivalent to less radiation dose, because of not only a narrower beam but also a more relaxed angular sampling requirement in the longitudinal studies or multi-scale scenarios. Reduced Angular Sampling Rate Need 4 Projections Need 2 Projections Statistical Interior Tomography Phantom [0.9 1.1] ROI 200,000 photons ITHT Known sub-Region Work in progress from Qiong Xu & Xuanqin Mou (China) in collaboration with Wang G & Yu HY 50,000 photons Hilbert Region ML Outline • Less Is Deeper • Less Is Larger • Less Is Faster • Less Is Less • Less Is More Less Is More Use of less data is advantageous in more modalities beyond CT, such as other straight-ray tomographic techniques and even in small-angle curvilinear geometry, and more applications of various types. Furthermore, less data means more computational time! Interior-MRI …………………………………… Interior-MRI Traditional MRI Interior MRI Zhang J, Yu HY, Corum C, Garwood M, Wang G: Exact and stable interior ROI reconstruction for radial MRI. SPIE 7258: 2585G, 8 pages, Feb. 2009, Orlando, FL, USA Interior Electron Tomography Ge Wang, Hengyong Yu Limited Angle Interior Tomography Interior SPECT Interior-SPECT FOV µa =0.3cm-1 Support Yu HY, Yang JS, Jiang M, Wang G: Interior SPECT- Exact and stable ROI reconstruction from uniformly attenuated local projections; Communications in Numerical Methods in Engineering, 25(6):693-710, 2009 Noisy data Known µa =0.15cm-1 Ideal data µa =0cm-1 Practical Implications Conclusion • Less Is Deeper • Less Is Larger • Less Is Faster • Less Is Less • Less Is More Less Is Not Always Better Link of Localities Pictures from http://www.bing.com Future Work • Multi-scale Interior Tomography • Multi-parameter Interior Tomography • Multi-energy Interior Tomography SBES Advanced Multi-scale CT Facility Smaller Scales? Larger Scales? Multi-parameter CT Grating-based Imaging Dark-field Tomography Wang G, Cong W, Shen H, Zou Y: Varying Collimation for Dark-Field Extraction. International Journal of Biomedical Imaging. 2009, Article ID 847537, 2010 Multi-energy CT Future Work • Theoretical Extension • Computational Optimization • Systematic Evaluation • Biomedical Applications • Interdisciplinary Collaboration Acknowledgment The results in this presentation are of collaborative nature. Major collaborators include Drs. Hengyong Yu, Yangbo Ye, Jiangsheng Yang, Ming Jiang, Steve Wang, Michael Fesser, Erik Ritman, Deepak Bharkhada, Bruno DeMan, Guohua Cao, Otto Zhou, Alexander Katsevich, et al. The work was partially supported by National Institutes of Health/National Institute of Biomedical Imaging and Bioengineering Grants EB002667, EB009275, and EB011785 as well as National Science Foundation NSF/CMMI 0923297. Thank You!