层析合成
成像体模
计算机科学
乳腺摄影术
迭代重建
计算机视觉
投影(关系代数)
人工智能
断层重建
光学
物理
乳腺癌
医学
算法
癌症
内科学
作者
Yiheng Zhang,Heang‐Ping Chan,Berkman Sahiner,Jun Wei,Mitchell M. Goodsitt,Lubomir M. Hadjiiski,Jun Ge,Chuan Zhou
摘要
Digital tomosynthesis mammography (DTM) is a promising approach to breast cancer detection. DTM can provide 3D structural information of the breast tissue by reconstructing the imaged volume from 2D projections acquired at different angles in a limited angular range. In this work, we investigate the application of the Simultaneous Algebraic Reconstruction Technique (SART) to this limited-angle cone-beam tomographic problem. A second generation GE prototype tomosynthesis mammography system was used in this study. Projection-view images of different breast phantoms were acquired from 21 angles in 3° increments over a ±30° angular range. The digital detector is stationary during image acquisition. We used an ACR phantom and two additional phantoms to evaluate the image quality and reconstruction artifacts. The Back-Projection (BP) method was also implemented for comparison to SART. The contrast-to-noise ratio (CNR), line profile of features and an artifact spread function (ASF) were used to quantitatively evaluate the reconstruction results. Preliminary results show that both BP and SART can separate superimposed phantom structures along the Z direction, but SART is more effective in improving the conspicuity of tissue-mimicking details and suppressing interplane blurring. For the phantoms with homogeneous background, the BP method resulted in less noisy reconstruction and higher CNR values for masses than SART, but SART provided greater enhancement in the contrast of calcification clusters and the edge sharpness of masses and fibrils. It was shown that acceptable reconstruction can be achieved by SART after only one iteration.
科研通智能强力驱动
Strongly Powered by AbleSci AI