This paper describes a universal approach to automatic segmentation of different internal organ and tissue regions in three-dimensional (3D) computerized tomography (CT) scans. The proposed approach combines object localization, a probabilistic atlas, and 3D GrabCut techniques to achieve automatic and quick segmentation. The proposed method first detects a tight 3D bounding box that contains the target organ region in CT images and then estimates the prior of each pixel inside the bounding box belonging to the organ region or background based on a dynamically generated probabilistic atlas. Finally, the target organ region is separated from the background by using an improved 3D GrabCut algorithm. A machine-learning method is used to train a detector to localize the 3D bounding box of the target organ using template matching on a selected feature space. A content-based image retrieval method is used for online generation of a patient-specific probabilistic atlas for the target organ based on a database. A 3D GrabCut algorithm is used for final organ segmentation by iteratively estimating the CT number distributions of the target organ and backgrounds using a graph-cuts algorithm. We applied this approach to localize and segment twelve major organ and tissue regions independently based on a database that includes 1300 torso CT scans. In our experiments, we randomly selected numerous CT scans and manually input nine principal types of inner organ regions for performance evaluation. Preliminary results showed the feasibility and efficiency of the proposed approach for addressing automatic organ segmentation issues on CT images.
The multi-resolution technology was developed for dynamic flat panel detectors for X-ray imaging. This multi-resolution
technology allows us to switch between the 1 x 1 mode (150 μm square) and the 2 x 2 mode (300 μm square)
instantaneously, using an external control. We developed a 17" x 17" dynamic detector using this multi-resolution
technology. This novel dynamic detector has a high sensitivity and a high-speed readout and it can reduce the radiation
exposure dose and deliver a smooth image. The key feature of our multi-resolution technology is capable of reading 4
pixel signals simultaneously. The sensitivity and the readout speed in the 2 x 2 mode were 4 times higher than those in
the 1 x 1 mode. The multi-resolution technology was implemented using a unique thin film transistor structure; that is,
one pixel has two switches, each of which are turned on/off depending on the readout mode. As a result, the dynamic
detector with a large active area of 17" x 17" realized a high detective quantum efficiency value of 40% under the low
radiation of RQA5 20 nGy and a high-speed readout of 30 frames/sec. This multi-resolution technology made it possible
to reduce the radiation exposure dose in a variety of applications.
A novel amorphous selenium (a-Se) detector with the hexagonal pixel has been developed for full-field digital
mammography. The pixel area of the detector was designed to be same as that of the 68 μm square pixel detector, while
the pixel pitch between neighboring pixels was set to be 73-75 μm in six directions. By applying the hexagonal pixels,
the sensitivity of the detector has improved by 18% compared with a conventional square pixel. A simulation showed
that the hexagonal pixel provided a more uniform electric field in the a-Se layer than the square pixel, which has lead to
higher sensitivity. The modulation transfer function of the detector was 92 % at 2 mm<sup>-1</sup> and 62 % at 5 mm<sup>-1</sup>. These
values were as high as that of a conventional a-Se detector with 50 μm square pixels. As a result, the detective quantum
efficiency of this detector achieved 75 % with 5 mR and 72 % with 2.5mR at 2 mm<sup>-1</sup>. The exposure conditions were 28
kV W/Rh with a 2 mm aluminum filter. Therefore, the new detector can reduce the exposure dose while maintaining a
high image quality.