🤖 AI Summary
Addressing the challenges of modeling long-range spatial dependencies, high computational cost, and reliance on large-scale in-domain pretraining in multi-label abnormality classification from 3D chest CT volumes, this paper proposes a graph-structured spectral-domain learning method. We represent the CT volume as a structured graph, with axial slice triplets serving as nodes, and employ spectral graph convolution to efficiently capture global spatial relationships. By integrating anatomical priors, our approach overcomes the limited receptive field of 3D convolutions and the prohibitive computational overhead of Vision Transformers (ViTs), while eliminating the need for large-scale in-domain pretraining. Experiments demonstrate competitive performance on multi-label classification, strong cross-dataset generalization, and robustness to z-axis translation. Ablation studies validate the effectiveness of both the graph construction strategy and the spectral convolution design.
📝 Abstract
With the increasing number of CT scan examinations, there is a need for automated methods such as organ segmentation, anomaly detection and report generation to assist radiologists in managing their increasing workload. Multi-label classification of 3D CT scans remains a critical yet challenging task due to the complex spatial relationships within volumetric data and the variety of observed anomalies. Existing approaches based on 3D convolutional networks have limited abilities to model long-range dependencies while Vision Transformers suffer from high computational costs and often require extensive pre-training on large-scale datasets from the same domain to achieve competitive performance. In this work, we propose an alternative by introducing a new graph-based approach that models CT scans as structured graphs, leveraging axial slice triplets nodes processed through spectral domain convolution to enhance multi-label anomaly classification performance. Our method exhibits strong cross-dataset generalization, and competitive performance while achieving robustness to z-axis translation. An ablation study evaluates the contribution of each proposed component.