Segmentation-free Radon transform algorithm to detect orientation and size of tissue structures in multiphoton...

Area of Science:

• Biophysics
• Medical Imaging
• Computational Biology

Background:

• Understanding biological tissue structure is crucial for physiology and pathology.
• In vivo imaging, like multiphoton microscopy, offers high-resolution tissue visualization.
• Automated orientation analysis is difficult due to noise, complexity, and subjective manual annotations.

Purpose of the Study:

• To present a robust, annotation-free algorithm for structural orientation analysis.
• To improve objectivity and efficiency in analyzing multimodal imaging datasets.
• To avoid preprocessing artifacts in structural orientation analysis.

Main Methods:

• Utilized a patch-based Radon transform for detecting oriented structures in noisy images.
• Analyzed projection peaks in Radon space to enhance small structures and minimize noise.
• Evaluated the algorithm using synthetic and in vivo datasets, comparing with human annotations.

Main Results:

• Achieved strong agreement with human annotations, exceeding 88% detection accuracy.
• Demonstrated robustness across different imaging modalities.
• Highlighted the benefits of an objective, mathematical approach over subjective human ratings.

Conclusions:

• The algorithm offers a robust and adaptable solution for structural orientation analysis in biological images.
• Quantifies tissue component orientation without preprocessing artifacts.
• Valuable for high-resolution, dynamic studies in tissue architecture and biomechanics.