arXiv subjects

Ge Hu

Publications and source records attributed to Ge Hu.

4 recordsLinked to original sources

arXiv2026

DeepFAN, a transformer-based deep learning model for human-artificial intelligence collaborative assessment of incidental pulmonary nodules in CT scans: a multi-reader, multi-case trial

The widespread adoption of CT has notably increased the number of detected lung nodules. However, current deep learning methods for classifying benign and malignant nodules often fail to comprehensively integrate global and local features, and most of them have not been validated through clinical trials. To address this, we developed DeepFAN, a transformer-based model trained on over 10K pathology-confirmed nodules and further conducted a multi-reader, multi-case clinical trial to evaluate its efficacy in assisting junior radiologists. DeepFAN achieved diagnostic area under the curve (AUC) of 0.939 (95% CI 0.930-0.948) on an internal test set and 0.954 (95% CI 0.934-0.973) on the clinical trial dataset involving 400 cases across three independent medical institutions. Explainability analysis indicated higher contributions from global than local features. Twelve readers' average performance significantly improved by 10.9% (95% CI 8.3%-13.5%) in AUC, 10.0% (95% CI 8.9%-11.1%) in accuracy, 7.6% (95% CI 6.1%-9.2%) in sensitivity, and 12.6% (95% CI 10.9%-14.3%) in specificity (P<0.001 for all). Nodule-level inter-reader diagnostic consistency improved from fair to moderate (overall k: 0.313 vs. 0.421; P=0.019). In conclusion, DeepFAN effectively assisted junior radiologists and may help homogenize diagnostic quality and reduce unnecessary follow-up of indeterminate pulmonary nodules. Chinese Clinical Trial Registry: ChiCTR2400084624.

cs.CV↗

arXiv2025

Filter2Noise: A Framework for Interpretable and Zero-Shot Low-Dose CT Image Denoising

Noise in low-dose computed tomography (LDCT) can obscure important diagnostic details. While deep learning offers powerful denoising, supervised methods require impractical paired data, and self-supervised alternatives often use opaque, parameter-heavy networks that limit clinical trust. We propose Filter2Noise (F2N), a novel self-supervised framework for interpretable, zero-shot denoising from a single LDCT image. Instead of a black-box network, its core is an Attention-Guided Bilateral Filter, a transparent, content-aware mathematical operator. A lightweight attention module predicts spatially varying filter parameters, making the process transparent and allowing interactive radiologist control. To learn from a single image with correlated noise, we introduce a multi-scale self-supervised loss coupled with Euclidean Local Shuffle (ELS) to disrupt noise patterns while preserving anatomical integrity. On the Mayo Clinic LDCT Challenge, F2N achieves state-of-the-art results, outperforming competing zero-shot methods by up to 3.68 dB in PSNR. It accomplishes this with only 3.6k parameters, orders of magnitude fewer than competing models, which accelerates inference and simplifies deployment. By combining high performance with transparency, user control, and high parameter efficiency, F2N offers a trustworthy solution for LDCT enhancement. We further demonstrate its applicability by validating it on clinical photon-counting CT data. Code is available at: https://github.com/sypsyp97/Filter2Noise.

eess.IV↗

arXiv2020

Topological transition in monolayer blue phosphorene with transiton-metal adatom under stain

We carried out first-principles calculations to investigate the electronic properties of the monolayer blue phosphorene (BlueP) decorated by the group-IVB transition-metal adatoms (Cr, Mo and W), and found that the Cr-decorated BlueP is a magnetic half metal, while the Mo- and W-decorated BlueP are semiconductors with band gaps smaller than 0.2 eV. Compressive strains make the band gaps close and reopen and band inversions occur during this process, which induces topological transitions in the Mo-decorated BlueP (with strain of -5.75%) and W-decorated BlueP (with strain of -4.25%) from normal insulators to topological insulators (TIs). The TI gap is 94 meV for the Mo-decorated BlueP and 150~220 meV for the W-decorated BlueP. Our findings pave the way to engineer topological phases in the monolayer BlueP with transition-metal adatoms.

cond-mat.mtrl-sci↗

arXiv2018

Alloying-induced topological transition in 2D transition-metal dichalcogenide semiconductors

Research on two-dimensional (2D) topological insulators (TIs) is obstructed due to the lack of feasible approaches to growing 2D TIs in experiments. Through systematic first-principles calculations and tight-binding simulations, we propose that alloying Os in 2D MoX2 (X = S, Se, Te) monolayers is an effective approach to inducing semiconductor-to-TI transitions, with sizable nontrivial gaps of 25-37 meV. Analysis of the electronic structures reveals that the topological property mainly originates from the 5d orbitals of the Os atom. Furthermore, the TI gaps can be modulated by external biaxial strain.

cond-mat.mtrl-sci↗