Prior knowledge-based blood vessel reconstruction and thrombus visualization using confocal laser scanning microscopic images

This paper describes an algorithm for the reconstruction of 3D Blood Vessel (BV) structure and thrombus. Biologists use a Confocal Laser Scanning Microscope (CLSM) to observe the brain of a mouse. However, the CLSM images obtained are in 2D, which limits the understanding of the 3D structure. In a previous method, the 3D BV structure was reconstructed from the centerline of the BV extracted from CLSM images. However, the reconstructed 3D BV structure is not realistic due to the influence of image noise. Therefore, we apply a smooth operation to the centerline of BV to make the reconstructed 3D BV structure more realistic. Meanwhile, we segment the thrombus from the CLSM images and visualize it within the reconstructed 3D BV. The 3D BV structure reconstruction method consists of three steps: (1) BV region segmentation, (2) centerline calculation and 3D mapping, and (3) 3D BV structure reconstruction. The 3D thrombus is reconstructed based on the reconstructed BV shape. Experimental results showed that the BVs obtained by the proposed method were smoother and more realistic compared to the previous method. In addition, the proposed method displayed the reconstructed thrombus in 3D space together with the BV, allowing 3D viewing from any viewpoint.