🫁 NeuroScan AI

Description: Comparison of real lung CT scans from the Learn2Reg Challenge. The baseline shows the lungs at end-inspiration (maximum volume), while the followup shows end-expiration (minimum volume). The significant anatomical deformation is visible, especially around the diaphragm.

Description: Registration test using real breathing motion data. The algorithm attempts to align the inspiration and expiration scans.

Results:
• MAE Before Registration: 432.47 HU
• MAE After Registration: 385.49 HU
• Improvement: 10.9%

Note: Large residual differences around the diaphragm are expected due to the significant breathing motion (typically 2-4 cm displacement).

Description: Three orthogonal views (axial, coronal, sagittal) showing the lung anatomy during inspiration and expiration phases.

Description: Quantitative analysis of lung volume change during breathing. The cyan contour represents inspiration (larger volume), while the green contour represents expiration (smaller volume). This demonstrates the physiological lung volume change that our registration algorithm must handle.