Medical Image Alignment for Different Resolutions and Fields of View Using Contrastive Learning with Feature-Level...

Area of Science:

• Computer Vision
• Image Processing
• Machine Learning

Background:

• Image alignment is crucial for comparing and merging images.
• Conventional methods struggle with variations in scale and image modality.
• Noise and scale differences present significant challenges in image alignment.

Purpose of the Study:

• To develop a novel feature-space-based alignment method.
• To enable effective image alignment across different scales and modalities.
• To improve modality-independent and noise-resilient similarity calculations.

Main Methods:

• Utilizing a feature-space-based approach for image alignment.
• Incorporating contrastive learning to enhance feature representation.
• Developing discriminative features for improved alignment accuracy.

Main Results:

• The proposed method achieves effective alignment for images with varying scales.
• The alignment process is robust against noise.
• Demonstrated superior performance in simulated real-world conditions.

Conclusions:

• The feature-space-based alignment method offers a robust solution for cross-scale and cross-modality image alignment.
• Contrastive learning significantly enhances feature representation for alignment tasks.
• The method shows promise for various applications requiring accurate image superposition.