What is CT reconstruction?

In computed tomography, reconstruction is the core computational process of converting hundreds or thousands of 2D transmission images, known as projection data, into a 3D volumetric dataset that represents the internal structure of the scanned object.

During a CT scan, X-rays are passed through the sample as it is incrementally rotated, typically over a full 360 degrees. Each projection image captured is like a conventional X-ray radiograph-a "shadow image" where internal structures are superimposed. From these 2D images alone, it is impossible to know the precise location and geometry of internal features.

This is where reconstruction comes in. The computer uses all the projection images from every angle as input. Using sophisticated mathematical algorithms, most commonly a version of Filtered Back Projection (FBP) such as the Feldkamp-Davis-Kress (FDK) algorithm, the computer reconstructs the 3D volume by back-projecting these filtered "shadows" into a 3D space. This is a highly advanced process, analogous to recreating a complex 3D object using only its shadows from multiple light sources.

The result of this reconstruction is a 3D volume composed of a vast number of voxels (volumetric pixels). Once this 3D data is generated, we can perform various analyses, such as virtually slicing the object to view any cross-section, rendering it as a 3D model, or taking precise internal measurements.

Matsusada Precision optimizes its reconstruction software with the latest GPU (Graphics Processing Unit) parallel computing technology. This allows our systems to complete this once time-consuming process in a remarkably short time, significantly improving the efficiency of your inspection and analysis workflows.

Related FAQ

• Does operating the system require specialized knowledge? Can a novice use it?