Search papers, labs, and topics across Lattice.
This paper introduces Perturb-and-Restore (P&R), a simulation-driven augmentation framework to address data imbalance in detecting structural chromosomal abnormalities. P&R generates synthetic abnormal chromosomes by perturbing normal chromosome banding patterns and then using a diffusion network to reconstruct realistic content and edges. An energy-guided sampling strategy prioritizes high-quality synthetic samples based on the energy distribution of real samples, leading to state-of-the-art performance on a new dataset of 260,000 chromosome images.
A diffusion model can generate high-quality synthetic chromosome images, boosting anomaly detection by nearly 14% F1 score and reducing reliance on scarce real-world abnormal samples.
Detecting structural chromosomal abnormalities is crucial for accurate diagnosis and management of genetic disorders. However, collecting sufficient structural abnormality data is extremely challenging and costly in clinical practice, and not all abnormal types can be readily collected. As a result, deep learning approaches face significant performance degradation due to the severe imbalance and scarcity of abnormal chromosome data. To address this challenge, we propose a Perturb-and-Restore (P&R), a simulation-driven structural augmentation framework that effectively alleviates data imbalance in chromosome anomaly detection. The P&R framework comprises two key components: (1) Structure Perturbation and Restoration Simulation, which generates synthetic abnormal chromosomes by perturbing chromosomal banding patterns of normal chromosomes followed by a restoration diffusion network that reconstructs continuous chromosome content and edges, thus eliminating reliance on rare abnormal samples; and (2) Energy-guided Adaptive Sampling, an energy score-based online selection strategy that dynamically prioritizes high-quality synthetic samples by referencing the energy distribution of real samples. To evaluate our method, we construct a comprehensive structural anomaly dataset consisting of over 260,000 chromosome images, including 4,242 abnormal samples spanning 24 categories. Experimental results demonstrate that the P&R framework achieves state-of-the-art (SOTA) performance, surpassing existing methods with an average improvement of 8.92% in sensitivity, 8.89% in precision, and 13.79% in F1-score across all categories.
