The marching cubes algorithm is a high-resolution 3D surface construction method that generates triangle models of constant density surfaces from 3D medical data. It uses a divide-and-conquer approach to determine inter-slice connectivity and creates a case table to define triangle topology. The algorithm processes 3D medical data in scan-line order, calculating triangle vertices using linear interpolation. It also computes the gradient of the original data, normalizes it, and uses it for shading. The algorithm maintains inter-slice connectivity, surface data, and gradient information to produce detailed images. Results from CT, MR, and SPECT data demonstrate the algorithm's quality and functionality. Improvements reduce processing time and add solid modeling capabilities. The algorithm involves four steps: data acquisition, image processing, surface construction, and display. Data acquisition involves sampling properties in a patient to produce 2D slices. Image processing may be used to find structures or filter data. Surface construction creates a surface model from 3D data, with users specifying a density value. Display involves rendering the surface using techniques like ray casting, depth shading, and color shading. Several approaches exist for 3D surface generation, but they often lack detail or introduce artifacts. The marching cubes algorithm uses information from the original 3D data to derive inter-slice connectivity, surface location, and surface gradient. The resulting triangle model can be displayed using standard rendering algorithms. The marching cubes algorithm involves two primary steps: locating the surface corresponding to a user-specified value and creating triangles, then calculating normals for each triangle vertex. It uses a divide-and-conquer approach to locate the surface in a logical cube created from eight pixels. The algorithm determines how the surface intersects the cube and moves to the next cube. It assigns a value to each cube's vertex based on whether it exceeds the surface value. The surface intersects edges where one vertex is inside and the other is outside. The algorithm creates a table to look up surface-edge intersections based on the cube's vertex labeling. The algorithm reduces the number of cases from 256 to 14 by using symmetries of the cube. It calculates the unit normal at each cube vertex using central differences and interpolates the normal to each triangle vertex. The final step outputs the triangle vertices and vertex normals. Efficiency enhancements allow the algorithm to take advantage of pixel-to-pixel, line-to-line, and slice-to-slice coherence. Functional enhancements add solid modeling capabilities, including Boolean operations for cutting and capping solid models. Texture mapping is also implemented by finding triangles on a plane's surface and attenuating the normal's length using original slice data. The algorithm is implemented in C and runs on various systems, including Sun Workstations, VAX, and IBM 3081. It displays models using an in-house z-buffer program or a General Electric Graphicon 700. The algorithm can handle a wide range of data, including CT, MR, and SPECT dataThe marching cubes algorithm is a high-resolution 3D surface construction method that generates triangle models of constant density surfaces from 3D medical data. It uses a divide-and-conquer approach to determine inter-slice connectivity and creates a case table to define triangle topology. The algorithm processes 3D medical data in scan-line order, calculating triangle vertices using linear interpolation. It also computes the gradient of the original data, normalizes it, and uses it for shading. The algorithm maintains inter-slice connectivity, surface data, and gradient information to produce detailed images. Results from CT, MR, and SPECT data demonstrate the algorithm's quality and functionality. Improvements reduce processing time and add solid modeling capabilities. The algorithm involves four steps: data acquisition, image processing, surface construction, and display. Data acquisition involves sampling properties in a patient to produce 2D slices. Image processing may be used to find structures or filter data. Surface construction creates a surface model from 3D data, with users specifying a density value. Display involves rendering the surface using techniques like ray casting, depth shading, and color shading. Several approaches exist for 3D surface generation, but they often lack detail or introduce artifacts. The marching cubes algorithm uses information from the original 3D data to derive inter-slice connectivity, surface location, and surface gradient. The resulting triangle model can be displayed using standard rendering algorithms. The marching cubes algorithm involves two primary steps: locating the surface corresponding to a user-specified value and creating triangles, then calculating normals for each triangle vertex. It uses a divide-and-conquer approach to locate the surface in a logical cube created from eight pixels. The algorithm determines how the surface intersects the cube and moves to the next cube. It assigns a value to each cube's vertex based on whether it exceeds the surface value. The surface intersects edges where one vertex is inside and the other is outside. The algorithm creates a table to look up surface-edge intersections based on the cube's vertex labeling. The algorithm reduces the number of cases from 256 to 14 by using symmetries of the cube. It calculates the unit normal at each cube vertex using central differences and interpolates the normal to each triangle vertex. The final step outputs the triangle vertices and vertex normals. Efficiency enhancements allow the algorithm to take advantage of pixel-to-pixel, line-to-line, and slice-to-slice coherence. Functional enhancements add solid modeling capabilities, including Boolean operations for cutting and capping solid models. Texture mapping is also implemented by finding triangles on a plane's surface and attenuating the normal's length using original slice data. The algorithm is implemented in C and runs on various systems, including Sun Workstations, VAX, and IBM 3081. It displays models using an in-house z-buffer program or a General Electric Graphicon 700. The algorithm can handle a wide range of data, including CT, MR, and SPECT data