No Image, No Problem: End-to-End Multi-Task Cardiac Analysis from Undersampled k-Space
2026-03-10 • Computer Vision and Pattern Recognition
Computer Vision and Pattern RecognitionArtificial Intelligence
AI summaryⓘ
The authors explain that traditional heart MRI methods first create images from raw signals and then analyze them, which can cause problems like missing details. Instead, they propose a new method called k-MTR that directly learns useful medical information from the raw MRI signals without making an image first. By training on many simulated cases, their method captures important heart features in a simpler form that computers can use for diagnosing conditions. Their tests show it works well for tasks like recognizing diseases and identifying heart parts, making MRI analysis faster and potentially more accurate.
Cardiac MRIk-spaceUndersamplingRepresentation learningLatent spacePhenotype regressionDisease classificationAnatomical segmentationInverse problemMulti-task learning
Authors
Yundi Zhang, Sevgi Gokce Kafali, Niklas Bubeck, Daniel Rueckert, Jiazhen Pan
Abstract
Conventional clinical CMR pipelines rely on a sequential "reconstruct-then-analyze" paradigm, forcing an ill-posed intermediate step that introduces avoidable artifacts and information bottlenecks. This creates a fundamental mathematical paradox: it attempts to recover high-dimensional pixel arrays (i.e., images) from undersampled k-space, rather than directly extracting the low-dimensional physiological labels actually required for diagnosis. To unlock the direct diagnostic potential of k-space, we propose k-MTR (k-space Multi-Task Representation), a k-space representation learning framework that aligns undersampled k-space data and fully-sampled images into a shared semantic manifold. Leveraging a large-scale controlled simulation of 42,000 subjects, k-MTR forces the k-space encoder to restore anatomical information lost to undersampling directly within the latent space, bypassing the explicit inverse problem for downstream analysis. We demonstrate that this latent alignment enables the dense latent space embedded with high-level physiological semantics directly from undersampled frequencies. Across continuous phenotype regression, disease classification, and anatomical segmentation, k-MTR achieves highly competitive performance against state-of-the-art image-domain baselines. By showcasing that precise spatial geometries and multi-task features can be successfully recovered directly from the k-space representations, k-MTR provides a robust architectural blueprint for task-aware cardiac MRI workflows.