PET and MRI image fusion with multi-scale transform

سال انتشار: 1400
نوع سند: مقاله کنفرانسی
زبان: انگلیسی
مشاهده: 367

فایل این مقاله در 15 صفحه با فرمت PDF قابل دریافت می باشد

استخراج به نرم افزارهای پژوهشی:

لینک ثابت به این مقاله:

شناسه ملی سند علمی:

ENPMCONF05_030

تاریخ نمایه سازی: 10 اسفند 1400

چکیده مقاله:

Introduction: Generally, medical imaging is divided into structural and functional systems. Structural image such as MRI (magnetic resonance imaging) and CT (computed tomography) provide high-resolution images with anatomical information; functional image such as PET (positron emission tomography) and SPECT (single-photon emission computed tomography) provide functional information with low spatial resolution. A single image can’t fundamentally satisfy clinical needs, so combining anatomical and functional images provide useful information. Methods: In this paper, the NSCT (non-sub sampled Contourlet transform) was inspired by the non-subsampled wavelet transform or the stationary wavelet transform, which were computed with the à trous algorithm. To retain the directional and multiscale properties of the transform, the Laplacian Pyramid was replaced with a non-subsampled pyramid structure to retain the multiscale property, and a non-subsampled directional filter bank for directionality. Results: The first major notable difference is that up sampling and down sampling are removed from both processes, instead; the filters in both the Laplacian Pyramid, and the directional filter banks are up-sampled. In this algorithm, all of the high frequency and low frequency components are fussed by using fusion rules. Then the fused image is reconstructed by using the inverse NSCT with all coefficients. Conclusions: Simulation results show the proposed framework provides the effective fusion. Evaluation criteria are defined used to compare the multiscale and multiresolution representations for image fusion. Three different fusion rules are utilized for this purpose. The results show an absolute improvement of (۰.۱%) in mutual information over the existing methods.

نویسندگان

Leila Rahimi

Faculty of electrical and computer engineering, University of Tabriz

Sebelan Danishvar

Faculty of electrical and computer engineering, University of Tabriz