ConvFormer3D-TAP: Phase/Uncertainty-Aware Front-End Fusion for Cine CMR View Classification Pipelines

2026-04-13

AI Summary

This work addresses the challenge of inaccurate automatic classification of standard-view clinical cine cardiac MRI due to variations in imaging equipment, acquisition protocols, motion artifacts, and slice orientations. To this end, the authors propose ConvFormer3D-TAP, a novel architecture that integrates 3D convolutional tokenization with multi-scale self-attention, jointly modeling local anatomical structures and dynamic features across the entire cardiac cycle. The method innovatively combines convolutional inductive bias with hierarchical attention and employs a phase-aware, uncertainty-guided early fusion strategy enhanced by masked spatiotemporal reconstruction pretraining and uncertainty-weighted multi-segment aggregation. Evaluated on 150,974 clinical sequences, the approach achieves 96% accuracy, F1 scores ≥0.94 across all classes, and excellent calibration (ECE=0.025, Brier=0.040), substantially outperforming existing methods, particularly under ambiguous temporal phases and real-world clinical variability.

Technology Category

• Computer Vision: Multi-modal Vision
• Machine Learning: Calibration & Uncertainty Quantification
• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning

Application Category

• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Responsible Web: Machine-in-the-loop, human agency and autonomy
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

Reliable recognition of standard cine cardiac MRI views is essential because each view determines which cardiac anatomy is visualized and which quantitative analyses can be performed. Incorrect view identification, whether by a human reader or an automated deep learning system, can propagate errors into segmentation, volumetric assessment, strain analysis, and valve evaluation. However, accurate view classification remains challenging under routine clinical variability in scanner vendor, acquisition protocol, motion artifacts, and plane prescription. We present ConvFormer3D-TAP, a cine-specific spatiotemporal architecture that integrates 3D convolutional tokenization with multiscale self-attention. The model is trained using masked spatiotemporal reconstruction and uncertainty-weighted multi-clip fusion to enhance robustness across cardiac phases and ambiguous temporal segments. The design captures complementary cues: local anatomical structure through convolutional priors and long-range cardiac-cycle dynamics through hierarchical attention. On a cohort of 150,974 clinically acquired cine sequences spanning six standard cine cardiac MRI views, ConvFormer3D-TAP achieved 96% validation accuracy with per-class F1-scores >= 0.94 and strong calibration (ECE = 0.025; Brier = 0.040). Error analysis shows that residual confusions are concentrated in anatomically adjacent long-axis and LVOT/AV view pairs, consistent with intrinsic prescription overlap. These results support ConvFormer3D-TAP as a scalable front-end for view routing, filtering and quality control in end-to-end cMRI workflows.

Problem

• cine cardiac MRI
• view classification
• clinical variability
• motion artifacts
• plane prescription

Innovation

• ConvFormer3D-TAP
• spatiotemporal fusion
• uncertainty-aware classification
• masked reconstruction
• cardiac MRI view recognition