Searcharxiv⌕ Search

arXiv subjects

Tian Liu

Publications and source records attributed to Tian Liu.

At least 55 records · Page 3Linked to original sources
arXiv2023

Crystallographic-dependent bilinear magnetoelectric resistance in a thin WTe$_2$ layer

The recently reported Bilinear Magnetoeletric Resistance (BMR) in novel materials with rich spin textures, such as bismuth selenide (Bi$_2$Se$_3$) and tungsten ditelluride (WTe$_2$), opens new possibilities for probing the spin textures via magneto-transport measurements. By its nature, the BMR effect is directly linked to the crystal symmetry of the materials and its spin texture. Therefore, understanding the crystallographic dependency of the effect is crucial. Here we report the observation of crystallographic-dependent BMR in thin WTe$_2$ layers and explore how it is linked to its spin textures. The linear response measured in first harmonic signals and the BMR measured in second harmonic signals are both studied under a wide range of magnitudes and directions of magnetic field, applied current and at different temperatures. We discover a three-fold symmetry contribution of the BMR when current is applied along the a-axis of the WTe$_2$ thin layer at 10 K, which is absent for when current is applied along the b-axis.

cond-mat.mes-hall↗
arXiv2023

Data-Driven Volumetric Image Generation from Surface Structures using a Patient-Specific Deep Leaning Model

The advent of computed tomography significantly improves patient health regarding diagnosis, prognosis, and treatment planning and verification. However, tomographic imaging escalates concomitant radiation doses to patients, inducing potential secondary cancer. We demonstrate the feasibility of a data-driven approach to synthesize volumetric images using patient surface images, which can be obtained from a zero-dose surface imaging system. This study includes 500 computed tomography (CT) image sets from 50 patients. Compared to the ground truth CT, the synthetic images result in the evaluation metric values of 26.9 Hounsfield units, 39.1dB, and 0.965 regarding the mean absolute error, peak signal-to-noise ratio, and structural similarity index measure. This approach provides a data integration solution that can potentially enable real-time imaging, which is free of radiation-induced risk and could be applied to image-guided medical procedures.

physics.med-ph↗
arXiv2023

Cycle-guided Denoising Diffusion Probability Model for 3D Cross-modality MRI Synthesis

This study aims to develop a novel Cycle-guided Denoising Diffusion Probability Model (CG-DDPM) for cross-modality MRI synthesis. The CG-DDPM deploys two DDPMs that condition each other to generate synthetic images from two different MRI pulse sequences. The two DDPMs exchange random latent noise in the reverse processes, which helps to regularize both DDPMs and generate matching images in two modalities. This improves image-to-image translation ac-curacy. We evaluated the CG-DDPM quantitatively using mean absolute error (MAE), multi-scale structural similarity index measure (MSSIM), and peak sig-nal-to-noise ratio (PSNR), as well as the network synthesis consistency, on the BraTS2020 dataset. Our proposed method showed high accuracy and reliable consistency for MRI synthesis. In addition, we compared the CG-DDPM with several other state-of-the-art networks and demonstrated statistically significant improvements in the image quality of synthetic MRIs. The proposed method enhances the capability of current multimodal MRI synthesis approaches, which could contribute to more accurate diagnosis and better treatment planning for patients by synthesizing additional MRI modalities.

eess.IV↗
arXiv2023

Deep Learning in MRI-guided Radiation Therapy: A Systematic Review

MRI-guided radiation therapy (MRgRT) offers a precise and adaptive approach to treatment planning. Deep learning applications which augment the capabilities of MRgRT are systematically reviewed. MRI-guided radiation therapy offers a precise, adaptive approach to treatment planning. Deep learning applications which augment the capabilities of MRgRT are systematically reviewed with emphasis placed on underlying methods. Studies are further categorized into the areas of segmentation, synthesis, radiomics, and real time MRI. Finally, clinical implications, current challenges, and future directions are discussed.

physics.med-ph↗
arXiv2023

Denoising Magnetic Resonance Spectroscopy (MRS) Data Using Stacked Autoencoder for Improving Signal-to-Noise Ratio and Speed of MRS

Background: Magnetic resonance spectroscopy (MRS) enables non-invasive detection and measurement of biochemicals and metabolites. However, MRS has low signal-to-noise ratio (SNR) when concentrations of metabolites are in the range of the million molars. Standard approach of using a high number of signal averaging (NSA) to achieve sufficient NSR comes at the cost of a long acquisition time. Purpose: We propose to use deep-learning approaches to denoise MRS data without increasing the NSA. Methods: The study was conducted using data collected from the brain spectroscopy phantom and human subjects. We utilized a stack auto-encoder (SAE) network to train deep learning models for denoising low NSA data (NSA = 1, 2, 4, 8, and 16) randomly truncated from high SNR data collected with high NSA (NSA=192) which were also used to obtain the ground truth. We applied both self-supervised and fully-supervised training approaches and compared their performance of denoising low NSA data based on improved SNRs. Results: With the SAE model, the SNR of low NSA data (NSA = 1) obtained from the phantom increased by 22.8% and the MSE decreased by 47.3%. For low NSA images of the human parietal and temporal lobes, the SNR increased by 43.8% and the MSE decreased by 68.8%. In all cases, the chemical shift of NAA in the denoised spectra closely matched with the high SNR spectra, suggesting no distortion to the spectra from denoising. Furthermore, the denoising performance of the SAE model was more effective in denoising spectra with higher noise levels. Conclusions: The reported SAE denoising method is a model-free approach to enhance the SNR of low NSA MRS data. With the denoising capability, it is possible to acquire MRS data with a few NSA, resulting in shorter scan times while maintaining adequate spectroscopic information for detecting and quantifying the metabolites of interest.

physics.med-ph↗
arXiv2023

CBCT-Based Synthetic CT Image Generation Using Conditional Denoising Diffusion Probabilistic Model

Background: Daily or weekly cone-beam computed tomography (CBCT) scans are commonly used for accurate patient positioning during the image-guided radiotherapy (IGRT) process, making it an ideal option for adaptive radiotherapy (ART) replanning. However, the presence of severe artifacts and inaccurate Hounsfield unit (HU) values prevent its use for quantitative applications such as organ segmentation and dose calculation. To enable the clinical practice of online ART, it is crucial to obtain CBCT scans with a quality comparable to that of a CT scan. Purpose: This work aims to develop a conditional diffusion model to perform image translation from the CBCT to the CT domain for the image quality improvement of CBCT. Methods: The proposed method is a conditional denoising diffusion probabilistic model (DDPM) that utilizes a time-embedded U-net architecture with residual and attention blocks to gradually transform standard Gaussian noise to the target CT distribution conditioned on the CBCT. The model was trained on deformed planning CT (dpCT) and CBCT image pairs, and its feasibility was verified in brain patient study and head-and-neck (H&N) patient study. The performance of the proposed algorithm was evaluated using mean absolute error (MAE), peak signal-to-noise ratio (PSNR) and normalized cross-correlation (NCC) metrics on generated synthetic CT (sCT) samples. The proposed method was also compared to four other diffusion model-based sCT generation methods. Conclusions: The proposed conditional DDPM method can generate sCT from CBCT with accurate HU numbers and reduced artifacts, enabling accurate CBCT-based organ segmentation and dose calculation for online ART.

physics.med-ph↗
arXiv2023

Deep Learning-based Multi-Organ CT Segmentation with Adversarial Data Augmentation

In this work, we propose an adversarial attack-based data augmentation method to improve the deep-learning-based segmentation algorithm for the delineation of Organs-At-Risk (OAR) in abdominal Computed Tomography (CT) to facilitate radiation therapy. We introduce Adversarial Feature Attack for Medical Image (AFA-MI) augmentation, which forces the segmentation network to learn out-of-distribution statistics and improve generalization and robustness to noises. AFA-MI augmentation consists of three steps: 1) generate adversarial noises by Fast Gradient Sign Method (FGSM) on the intermediate features of the segmentation network's encoder; 2) inject the generated adversarial noises into the network, intentionally compromising performance; 3) optimize the network with both clean and adversarial features. Experiments are conducted segmenting the heart, left and right kidney, liver, left and right lung, spinal cord, and stomach. We first evaluate the AFA-MI augmentation using nnUnet and TT-Vnet on the test data from a public abdominal dataset and an institutional dataset. In addition, we validate how AFA-MI affects the networks' robustness to the noisy data by evaluating the networks with added Gaussian noises of varying magnitudes to the institutional dataset. Network performance is quantitatively evaluated using Dice Similarity Coefficient (DSC) for volume-based accuracy. Also, Hausdorff Distance (HD) is applied for surface-based accuracy. On the public dataset, nnUnet with AFA-MI achieves DSC = 0.85 and HD = 6.16 millimeters (mm); and TT-Vnet achieves DSC = 0.86 and HD = 5.62 mm. AFA-MI augmentation further improves all contour accuracies up to 0.217 DSC score when tested on images with Gaussian noises. AFA-MI augmentation is therefore demonstrated to improve segmentation performance and robustness in CT multi-organ segmentation.

eess.IV↗
arXiv2023

Long distance magnon transport in the van der Waals antiferromagnet CrPS$_4$

We demonstrate the potential of van der Waals magnets for spintronic applications by reporting long-distance magnon spin transport in the electrically insulating antiferromagnet chromium thiophosphate (CrPS$_4$) with perpendicular magnetic anisotropy. We inject and detect magnon spins non-locally by Pt contacts and monitor the non-local resistance as a function of an in-plane magnetic field up to 7 Tesla. We observe a non-local resistance over distances up to at least a micron below the Neel temperature (T$_{\rm N}$ = 38 Kelvin) close to magnetic field strengths that saturate the sublattice magnetizations.

cond-mat.mes-hall↗
arXiv2022

A Retrospective Study on the Investigation of Potential Clinical Benefits of Online Adaptive Proton Therapy for Head and Neck Cancer

Online adaptive proton therapy (APT) is an ideal solution theoretically, which however is challenging to proton clinics. Although multiple groups have been endeavoring to develop online APT technology, there is a concern in the radiotherapy community about the necessity of online APT because of its unknown impact on treatment outcomes. Hence, we have performed a retrospective study to investigate the potential clinical effects of online APT for HN cancer patients in relative to the current offline APT via simulations. To mimic an online APT treatment course, we have recalculated and evaluated the actual dose of the current treatment plan on patient actual treatment anatomy captured by cone beam CT for each fraction. The cumulative dose of simulated online APT courses was compared to actual offline APT courses and the initially designed treatment plan dose. For patients 1 and 2, the simulated online ART course maintained a relatively higher CTV dose coverages than the offline ART course, particularly for CTV-Low, which led to an improvement of 2.66% and 4.52% in TCP of CTV-Low. For patients 3 and 4, with clinically comparable CTV dose coverages, the simulated online ART course achieved better OAR sparing than the offline ART course. The mean doses of right parotid and oral cavity were decreased from 29.52 Gy relative biological effectiveness (RBE) and 41.89 Gy RBE to 22.16 Gy RBE and 34.61 Gy RBE for patient 3, leading to a reduce of 1.67% and 3.40% in NTCP for the two organs. Compared to the current clinical practice, the retrospective study indicated that online APT tended to spare more normal tissues by achieving the clinical goal with merely half of the positional uncertainty margin. Future studies are needed to help identify the patients with large potential benefits prior to treatment to conserve scarce clinical resources.

physics.med-ph↗
arXiv2022

Deep Learning-based Protoacoustic Signal Denoising for Proton Range Verification

Objective: Proton therapy offers an advantageous dose distribution compared to the photon therapy, since it deposits most of the energy at the end of range, namely the Bragg peak (BP). Protoacoustic technique was developed to in vivo determine the BP locations. However, it requires large dose delivery to the tissue to obtain an averaged acoustic signal with a sufficient signal to noise ratio (SNR), which is not suitable in clinics. We propose a deep learning-based technique to acquire denoised acoustic signals and reduce BP range uncertainty with much lower doses. Approach: Three accelerometers were placed on the distal surface of a cylindrical polyethylene (PE) phantom to collect protoacoustic signals. In total 512 raw signals were collected at each device. Device-specific stack autoencoder (SAE) denoising models were trained to denoise the input signals, which were generated by averaging 1, 2, 4, 8, 16, or 32 raw signals. Both supervised and unsupervised learning training strategies were tested for comparison. Mean squared error (MSE), signal-to-noise ratio (SNR) and the Bragg peak (BP) range uncertainty were used for model evaluation. Main results: After SAE denoising, the MSE was substantially reduced, and the SNR was enhanced. Overall, the supervised SAEs outperformed the unsupervised SAEs in BP range verification. For the high accuracy detector, it achieved a BP range uncertainty of 0.20 +/- 3.44 mm by averaging over 8 raw signals, while for the other two low accuracy detectors, they achieved the BP uncertainty of 1.44 +/- 6.45 mm and -0.23 +/- 4.88 mm by averaging 16 raw signals, respectively. Significance: We have proposed a deep learning based denoising method to enhance the SNR of protoacoustic measurements and improve the accuracy in BP range verification, which greatly reduces the dose and time for potential clinical applications.

physics.med-ph↗
arXiv2022

Early in vivo Radiation Damage Quantification for Pediatric Craniospinal Irradiation Using Longitudinal MRI for Intensity Modulated Proton Therapy

Purpose: Proton vertebral body sparing craniospinal irradiation (VBS CSI) treats the thecal sac while avoiding the anterior vertebral bodies in effort to reduce myelosuppression and growth inhibition. However, robust treatment planning needs to compensate proton range uncertainty, contributing unwanted doses within the vertebral bodies. This work aims to develop an early in vivo radiation damage quantification method using longitudinal magnetic resonance (MR) scans to quantify dose effect during fractionated CSI. Materials and methods: Ten pediatric patients were enrolled in a prospective clinical trial of proton VBS CSI receiving 23.4-36 Gy. Monte Carlo robust planning was used with spinal clinical target volumes defined as the thecal sac and neural foramina. T1/T2-weighted MR scans were acquired before, during, and after treatments to detect transition from hematopoietic to less metabolically active fatty marrow. MR signal intensity histograms at each time point were analyzed and fitted by multi-Gaussian models to quantify radiation damages. Results: Fatty marrow filtration was observed on MR images as early as the fifth fraction of treatment. Maximum radiation-induced marrow damage occurred 40-50 days from the treatment start, followed by marrow regeneration. The mean damage ratios were 0.23, 0.41, 0.59, and 0.54 corresponding to 10, 20, 40, and 60 days from the treatment start. Conclusions: We demonstrated a non-invasive method to identify early vertebral marrow damage based on radiation-induced fatty marrow replacement. The proposed method can be potentially used to quantify the quality of CSI vertebral sparing to preserve metabolically active hematopoietic bone marrow.

physics.med-ph↗
arXiv2022

Deep learning-based Fast Volumetric Image Generation for Image-guided Proton FLASH Radiotherapy

Proton FLASH therapy leverages ultra-high dose-rate radiation to enhance the sparing of organs at risk without compromising tumor control probability. To prepare for the delivery of high doses to targets, we aim to develop a deep learning-based image-guide framework to enable fast volumetric image reconstruction for accurate target localization before FLSAH beam delivery. The proposed framework comprises four modules, including orthogonal kV x-ray projection acquisition, DL-based volumetric image generation, image quality analyses, and water equivalent thickness evaluation. We investigated volumetric image reconstruction using four kV projection pairs with different source angles. Thirty lung patients were identified from the institutional database, and each patient contains a four-dimensional computed tomography dataset with ten respiratory phases. The retrospective patient study indicated that the proposed framework could reconstruct patient volumetric anatomy, including tumors and organs at risk from orthogonal x-ray projections. Considering all evaluation metrics, the kV projections with source angles of 135 and 225 degrees yielded the optimal volumetric images. The proposed framework has been demonstrated to reconstruct volumetric images with accurate lesion locations from two orthogonal x-ray projections. The embedded WET module can be used to detect potential proton beam-specific patient anatomy variations. The framework can deliver fast volumetric image generation and can potentially guide treatment delivery systems for proton FLASH therapy.

physics.med-ph↗
arXiv2022

MRI-based Material Mass Density and Relative Stopping Power Estimation via Deep Learning for Proton Therapy

Magnetic Resonance Imaging (MRI) is increasingly incorporated into treatment planning, because of its superior soft tissue contrast used for tumor and soft tissue delineation versus computed tomography (CT). However, MRI cannot directly provide mass density or relative stopping power (RSP) maps required for proton radiotherapy dose calculation. To demonstrate the feasibility of MRI-only based mass density and RSP estimation using deep learning (DL) for proton radiotherapy. A DL-based framework was developed to discover underlying voxel-wise correlation between MR images and mass density and RSP. Five tissue substitute phantoms including skin, muscle, adipose, 45% hydroxyapatite (HA), and spongiosa bone were customized for MRI scanning based on material composition information from ICRP reports. Two animal tissue phantoms made of pig brain and liver were prepared for DL training. In the phantom study, two DL models were trained: one containing clinical T1 and T2 MRIs and another incorporating zero echo time (ZTE) MRIs as input. In the patient application study, two DL models were trained: one including T1 and T2 MRIs as input, and one incorporating synthetic dual-energy computed tomography (sDECT) images to provide bone tissue information. In the phantom study, DL model based on T1 and T2 MRI demonstrated higher accuracy mass density and RSP estimation in skin, muscle, adipose, brain, and liver with mean absolute percentage errors (MAPE) of 0.42%, 0.14%, 0.19%, 0.78% and 0.26% for mass density and 0.30%, 0.11%, 0.16%, 0.61% and 0.23% for RSP, respectively. DL model incorporating ZTE MRI improved the accuracy of mass density and RSP estimation in 45% HA and spongiosa bone with MAPE at 0.23% and 0.09% for mass density and 0.19% and 0.07% for RSP, respectively. Results show feasibility of MRI-only based mass density and RSP estimation for proton therapy treatment planning using DL method.

physics.med-ph↗
arXiv2022

Feasibility Study of Hybrid Inverse Planning with Transmission Beams and Single-energy Spread-out Bragg Peaks for Proton Flash Radiotherapy

Ultra-high dose rate (FLASH) proton planning with only transmission beams (TBs) has limitations in normal tissue sparing. The single-energy spread-out Bragg peaks (SESOBPs) of FLASH dose rate have been demonstrated feasible for proton FLASH planning. A hybrid inverse optimization method was developed to combine the TBs and SESOBPs (TB-SESOBP) for FLASH planning. The SESOBPs were generated from spreading out the BPs by pre-designed general bar ridge filters and placed at the central target by range shifters to obtain a uniform dose within the target. The SESOBPs and TBs were fully sampled field-by-field allowing automatic spot selection and weighting in the optimization process. The TB-SESOBP plans were validated in comparison with the TB only (TB-only) plans and the plans with the combination of TBs and BPs (TB-BP) regarding 3D dose and dose rate distributions for five lung cases. Comparing to the TB-only plans, the mean spinal cord D1.2cc drastically reduced 41%, the mean lung V7Gy and V7.4Gy moderately reduced by up to 17% and the target dose homogeneity slightly increased in the TB-SESOBP plans. Comparable dose homogeneity was achieved in both TB-SESOBP and TB-BP plans. Besides, prominent improvements were achieved in lung sparing for the cases of relatively large targets by the TB-SESOBP plans comparing to the TB-BP plans. The targets were fully covered with the FLASH dose rate in all the three plans. For the OARs, V40Gy/s = 100% was achieved by the TB-only plans while V40Gy/s > 85% was obtained by the other two plans. We have demonstrated that the hybrid TB-SESOBP planning was feasible to achieve FLASH dose rate for proton therapy. The hybrid TB-SESOBP planning has great potential in improving OAR sparing while maintaining high target dose homogeneity, and can be potentially implemented for adaptive radiotherapy.

physics.med-ph↗
arXiv2022

Landmark Tracking in Liver US images Using Cascade Convolutional Neural Networks with Long Short-Term Memory

This study proposed a deep learning-based tracking method for ultrasound (US) image-guided radiation therapy. The proposed cascade deep learning model is composed of an attention network, a mask region-based convolutional neural network (mask R-CNN), and a long short-term memory (LSTM) network. The attention network learns a mapping from a US image to a suspected area of landmark motion in order to reduce the search region. The mask R-CNN then produces multiple region-of-interest (ROI) proposals in the reduced region and identifies the proposed landmark via three network heads: bounding box regression, proposal classification, and landmark segmentation. The LSTM network models the temporal relationship among the successive image frames for bounding box regression and proposal classification. To consolidate the final proposal, a selection method is designed according to the similarities between sequential frames. The proposed method was tested on the liver US tracking datasets used in the Medical Image Computing and Computer Assisted Interventions (MICCAI) 2015 challenges, where the landmarks were annotated by three experienced observers to obtain their mean positions. Five-fold cross-validation on the 24 given US sequences with ground truths shows that the mean tracking error for all landmarks is 0.65+/-0.56 mm, and the errors of all landmarks are within 2 mm. We further tested the proposed model on 69 landmarks from the testing dataset that has a similar image pattern to the training pattern, resulting in a mean tracking error of 0.94+/-0.83 mm. Our experimental results have demonstrated the feasibility and accuracy of our proposed method in tracking liver anatomic landmarks using US images, providing a potential solution for real-time liver tracking for active motion management during radiation therapy.

cs.CV↗
arXiv2022

Deformable Image Registration using Unsupervised Deep Learning for CBCT-guided Abdominal Radiotherapy

CBCTs in image-guided radiotherapy provide crucial anatomy information for patient setup and plan evaluation. Longitudinal CBCT image registration could quantify the inter-fractional anatomic changes. The purpose of this study is to propose an unsupervised deep learning based CBCT-CBCT deformable image registration. The proposed deformable registration workflow consists of training and inference stages that share the same feed-forward path through a spatial transformation-based network (STN). The STN consists of a global generative adversarial network (GlobalGAN) and a local GAN (LocalGAN) to predict the coarse- and fine-scale motions, respectively. The network was trained by minimizing the image similarity loss and the deformable vector field (DVF) regularization loss without the supervision of ground truth DVFs. During the inference stage, patches of local DVF were predicted by the trained LocalGAN and fused to form a whole-image DVF. The local whole-image DVF was subsequently combined with the GlobalGAN generated DVF to obtain final DVF. The proposed method was evaluated using 100 fractional CBCTs from 20 abdominal cancer patients in the experiments and 105 fractional CBCTs from a cohort of 21 different abdominal cancer patients in a holdout test. Qualitatively, the registration results show great alignment between the deformed CBCT images and the target CBCT image. Quantitatively, the average target registration error (TRE) calculated on the fiducial markers and manually identified landmarks was 1.91+-1.11 mm. The average mean absolute error (MAE), normalized cross correlation (NCC) between the deformed CBCT and target CBCT were 33.42+-7.48 HU, 0.94+-0.04, respectively. This promising registration method could provide fast and accurate longitudinal CBCT alignment to facilitate inter-fractional anatomic changes analysis and prediction.

eess.IV↗
arXiv2022

Multimodal Imaging-based Material Mass Density Estimation for Proton Therapy Using Physics-Constrained Deep Learning

Mapping computed tomography (CT) number to material property dominates the proton range uncertainty. This work aims to develop a physics-constrained deep learning-based multimodal imaging (PDMI) framework to integrate physics, deep learning, magnetic resonance imaging (MRI), and advanced dual-energy CT (DECT) to derive accurate patient mass density maps. Seven tissue substitute MRI phantoms were used for PDMI-based material calibration. The training inputs are from MRI and twin-beam dual-energy images acquired at 120 kVp with gold and tin filters. The feasibility investigation included an empirical DECT correlation and four residual networks (ResNet) derived from different training inputs and strategies by the PDMI framework. PRN-MR-DE and RN-MR-DE denote ResNet trained with and without a physics constraint using MRI and DECT images. PRN-DE and RN-DE represent ResNet trained with and without a physics constraint using DECT-only images. For the tissue surrogate study, PRN-MR-DE, PRN-DE, and RN-MR-DE result in mean mass density errors: -0.72%, 2.62%, -3.58% for adipose; -0.03%, -0.61%, and -0.18% for muscle; -0.58%, -1.36%, and -4.86% for 45% HA bone. The retrospective patient study indicated that PRN-MR-DE predicted the densities of soft tissue and bone within expected intervals based on the literature survey, while PRN-DE generated large density deviations. The proposed PDMI framework can generate accurate mass density maps using MRI and DECT images. The physics-constrained training can further enhance model efficacy, making PRN-MR-DE outperform RN-MR-DE. The patient investigation also shows that the PDMI framework has the potential to improve proton range uncertainty with accurate patient mass density maps.

physics.med-ph↗
arXiv2022

Technical Report: Assisting Backdoor Federated Learning with Whole Population Knowledge Alignment

Due to the distributed nature of Federated Learning (FL), researchers have uncovered that FL is vulnerable to backdoor attacks, which aim at injecting a sub-task into the FL without corrupting the performance of the main task. Single-shot backdoor attack achieves high accuracy on both the main task and backdoor sub-task when injected at the FL model convergence. However, the early-injected single-shot backdoor attack is ineffective because: (1) the maximum backdoor effectiveness is not reached at injection because of the dilution effect from normal local updates; (2) the backdoor effect decreases quickly as the backdoor will be overwritten by the newcoming normal local updates. In this paper, we strengthen the early-injected single-shot backdoor attack utilizing FL model information leakage. We show that the FL convergence can be expedited if the client trains on a dataset that mimics the distribution and gradients of the whole population. Based on this observation, we proposed a two-phase backdoor attack, which includes a preliminary phase for the subsequent backdoor attack. In the preliminary phase, the attacker-controlled client first launches a whole population distribution inference attack and then trains on a locally crafted dataset that is aligned with both the gradient and inferred distribution. Benefiting from the preliminary phase, the later injected backdoor achieves better effectiveness as the backdoor effect will be less likely to be diluted by the normal model updates. Extensive experiments are conducted on MNIST dataset under various data heterogeneity settings to evaluate the effectiveness of the proposed backdoor attack. Results show that the proposed backdoor outperforms existing backdoor attacks in both success rate and longevity, even when defense mechanisms are in place.

cs.AI↗