Computed tomography image source identification by discriminating CT-scanner image reconstruction process.

Area of Science:

• Medical Imaging
• Computer Vision
• Signal Processing

Background:

• Computed Tomography (CT) imaging is crucial in diagnostics.
• Identifying the specific CT scanner used for an image is important for quality control and research.
• Current methods for CT scanner identification are limited.

Purpose of the Study:

• To develop a method for identifying the specific Computed Tomography (CT) scanner that produced a given CT image.
• To leverage the unique signatures left by CT scanner reconstruction processes on X-ray detector noise.

Main Methods:

• Analysis of how CT reconstruction algorithms modify intrinsic sensor noise.
• Definition of image features that capture the CT acquisition system's footprint.
• Training a Support Vector Machine (SVM) classifier using these features.

Main Results:

• The proposed method successfully discriminates between CT-Scanner systems based on noise modification patterns.
• Experiments demonstrated high accuracy in identifying the origin of CT images across multiple manufacturers and models.
• The defined image features effectively serve as a unique footprint for CT acquisition systems.

Conclusions:

• It is feasible to identify the specific CT scanner model used to acquire an image.
• The approach offers a novel method for CT scanner attribution based on image noise analysis.
• This technique has potential applications in medical image forensics and quality assurance.