Evaluation of different segmentation methods of X-ray micro computed tomography images

Area of Science:

• Pharmaceutical Sciences
• Image Analysis
• Computational Pharmaceutics

Background:

• In silico tools require validation with real pharmaceutical formulations.
• X-ray micro-computed tomography (X-ray micro-CT) provides detailed tablet imaging.
• Accurate image segmentation is crucial for reliable in silico predictions.

Purpose of the Study:

• To compare segmentation methods for X-ray micro-CT images of tablets.
• To optimize image processing for predicting tablet disintegration and dissolution.
• To reduce processing time while improving evaluation accuracy.

Main Methods:

• Utilized ImageJ open-source software for image processing.
• Applied various threshold algorithms and k-means clustering with different cluster numbers.
• Analyzed processing pathways based on recovery rates and their ratios.

Main Results:

• Identified optimal segmentation pathways for individual and combined tablet batches.
• Achieved recovery rates of approximately 100% with the best pathways.
• Confirmed image processing accuracy through visual perception.

Conclusions:

• Segmentation method optimization significantly improves tablet disintegration/dissolution prediction.
• The developed methods enhance both accuracy and processing speed.
• Validated in silico tools can be effectively supported by advanced image analysis techniques.