We developed a new type 3-Dimensional Internal Structure Microscope (3D-ISM) for the observation of internal structures of samples. The internal structure of a sample is obtained by cutting it into thin slices and observing the cutting side continuously while cutting. The position of the camera, as well as the sample position are fixed. Therefore there is no shift between the sections, and this system can obtain a true color image of the sample, which is a high resolution and a high-quality three-dimensional image compared with X-ray CT and MRI. After repeatedly slicing a sample, the digital image data from the sectional views is transferred to a computer, where 3-dimensional images of the internal structure of the sample are reconstructed. Using this system we analyzed many biological organisms. Just now, we are developing the system which builds in a fluorescent microscope. in this paper, We therefore observed fluorescent-stained samples by a fluorescence microscope installed in the 3D-ISM instrument are shown.

This article presents the outline of the device and the principle of the observation. This system, consisting of a slicing device, reflecting optical microscope, a CCD camera, a laser video recorder, a PC-based control system and an image processing workstation. FIG. 1 shows a schematic view of the cross secsional machine. Before slicing, the sample is embedded into either paraffin or OCT compound. Subsequently, the sample is positioned into a cross-sectioning machine for the slicing operation, which is performed using a cutting blade. The maximum size of the sample which can be observed with this device is ϕ 8x10mm. The thickness of the slices can be adjusted mechanically between 0.5-10 μ m. The cutting speed can range from 0.5 to 4 slices/ s. Each cross-section is viewed by employing a epi-fluoresence microscope and a CCD camera. The object lens used for the observation can use the x2-80. The resulting image is recorded on a laser video disc. After the slicing operation, the 3D sample image is reconstructed on the basis of volume rendering by using a Workstation. These observations can be completely done by full automation.