Three-dimensional image construction using multi axis multi directional processing method for 3DX Multi Image Micro CT

Three-dimensional diagnostic images can now be reconstructed from CT or MR images. In 2000, a limited cone beam X-ray CT for dental use (3DX Multi Image Micro CT, hereafter referred to as 3DX) was developed by Arai et al. ThreeDX has the advantages of small X-ray irradiation areas, high resolution, and low exposed doses. However, because there is a lot of noise interference in the 3DX data, the obtained 3-dimensional reconstructed images are inadequate. The aims of this investigation was to improve the 3-dimensional reconstructed images of 3DX. Firstly, gray level control was employed to reduce the noise on 3-dimensional image data. The amount of gray level correction was selected by comparing the threshold level of 3-dimensional image data with the average level of neighboring voxel. Then, 3-dimensional image data was decomposed to a set of 2-dimensional images from multiple directions. Edge detection, binary processing and line thinning were carried out to detect outlines. Finally, the obtained clear edges of the 2-dimensional image were reconstructed into 3-dimensional images. Adaptive gray level control made it possible to separate the caput mandibulae and fossa mandibularis. By reconstructing 3-dimensional image data from multiple directions, the target image can be interpolated with noise reduction and the edge detection. Our results demonstrate that 3-dimensional reconstructed images which are applicable for dental treatment can be obtained by the proposed method. A 3-dimensional image reconstruction method from 3-dimensional image data obtained by 3DX is proposed. As a result, it became possible to construct improved 3-dimensional reconstructed images from 3DX data. The reconstructed images obtained by the proposed method are widely applicable in dentistry. (author)