SearcharxivSearch

arXiv subjects

Guoqing Wu

Publications and source records attributed to Guoqing Wu.

12 recordsLinked to original sources

SAMPO-Path: Segmentation Intent-Aligned Preference Optimization for Pathology Foundation Model Segmentation

Foundation models have shown strong performance in multi-object segmentation with visual prompts, yet histopathology images remain challenging due to high cellular density, heterogeneity, and the gap between pixel-level supervision and clinical segmentation intent (e.g., selectively segmenting nuclei of a specific type). In practice, such intents are expressed through diverse and noisy prompts, causing prompt-intent misalignment and inconsistent predictions. We introduce SAMPO (Segmentation Anything Model with Preference Optimization), a preference-aligned fine-tuning framework that explicitly aligns pathology foundation models with clinical segmentation intent. SAMPO is the first to adapt Direct Preference Optimization (DPO) to pure vision foundation models, enabling accurate segmentation from minimal and imperfect prompts. The framework features three key components: (1) online prompt-centric preference mining to synthesize preference pairs across prompt qualities; (2) multi-mask preference learning to leverage output ambiguity for fine-grained ranking supervision; and (3) a hybrid loss combining preference optimization with pixel-level supervision for stable training. Trained on two datasets covering four tasks and evaluated on corresponding test sets and 12 external validation datasets, SAMPO consistently improves segmentation accuracy, robustness to prompt variations, and clinical intent adherence in dense histopathology images.

cs.CV

MindShot: Multi-Shot Video Reconstruction from fMRI with LLM Decoding

Reconstructing dynamic videos from fMRI is important for understanding visual cognition and enabling vivid brain-computer interfaces. However, current methods are critically limited to single-shot clips, failing to address the multi-shot nature of real-world experiences. Multi-shot reconstruction faces fundamental challenges: fMRI signal mixing across shots, the temporal resolution mismatch between fMRI and video obscuring rapid scene changes, and the lack of dedicated multi-shot fMRI-video datasets. To overcome these limitations, we propose a novel divide-and-decode framework for multi-shot fMRI video reconstruction. Our core innovations are: (1) A shot boundary predictor module explicitly decomposing mixed fMRI signals into shot-specific segments. (2) Generative keyframe captioning using LLMs, which decodes robust textual descriptions from each segment, overcoming temporal blur by leveraging high-level semantics. (3) Novel large-scale data synthesis (20k samples) from existing datasets. Experimental results demonstrate our framework outperforms state-of-the-art methods in multi-shot reconstruction fidelity. Ablation studies confirm the critical role of fMRI decomposition and semantic captioning, with decomposition significantly improving decoded caption CLIP similarity by 71.8%. This work establishes a new paradigm for multi-shot fMRI reconstruction, enabling accurate recovery of complex visual narratives through explicit decomposition and semantic prompting.

cs.CV

A Heterogeneous Graph Neural Network Fusing Functional and Structural Connectivity for MCI Diagnosis

Brain connectivity alternations associated with brain disorders have been widely reported in resting-state functional imaging (rs-fMRI) and diffusion tensor imaging (DTI). While many dual-modal fusion methods based on graph neural networks (GNNs) have been proposed, they generally follow homogenous fusion ways ignoring rich heterogeneity of dual-modal information. To address this issue, we propose a novel method that integrates functional and structural connectivity based on heterogeneous graph neural networks (HGNNs) to better leverage the rich heterogeneity in dual-modal images. We firstly use blood oxygen level dependency and whiter matter structure information provided by rs-fMRI and DTI to establish homo-meta-path, capturing node relationships within the same modality. At the same time, we propose to establish hetero-meta-path based on structure-function coupling and brain community searching to capture relations among cross-modal nodes. Secondly, we further introduce a heterogeneous graph pooling strategy that automatically balances homo- and hetero-meta-path, effectively leveraging heterogeneous information and preventing feature confusion after pooling. Thirdly, based on the flexibility of heterogeneous graphs, we propose a heterogeneous graph data augmentation approach that can conveniently address the sample imbalance issue commonly seen in clinical diagnosis. We evaluate our method on ADNI-3 dataset for mild cognitive impairment (MCI) diagnosis. Experimental results indicate the proposed method is effective and superior to other algorithms, with a mean classification accuracy of 93.3%.

cs.CV

Observation of orbital ordering and origin of the nematic order in FeSe

To elucidate the origin of nematic order in FeSe, we performed field-dependent 77Se-NMR measurements on single crystals of FeSe. We observed orbital ordering from the splitting of the NMR spectra and Knight shift and a suppression of it with magnetic field B0 up to 16 T applied parallel to the Fe-planes. There is a significant change in the distribution and magnitude of the internal magnetic field across the orbital ordering temperature Torb while stripe-type antiferromagnetism is absent. Giant antiferromagnetic (AFM) spin fluctuations measured by the NMR spin-lattice relaxation are gradually developed starting at ~ 40 K, which is far below the nematic ordering temperature Tnem. These results demonstrate that orbital ordering is the origin of the nematic order, and the AFM spin fluctuation is the driving mechanism of superconductivity in FeSe under the presence of the nematic order.

cond-mat.supr-con

Investigation of the magnetic dipole field at the atomic scale in quasi-one-dimensional paramagnetic conductor Li$_{0.9}$Mo$_{6}$O$_{17}$

We report magnetic dipole field investigation at the atomic scale in a single crystal of quasi-one-dimensional (Q1D) paramagnetic conductor Li$_{0.9}$Mo$_{6}$O$_{17}$, using a paramagnetic electron model and $^{7}$Li-NMR spectroscopy measurements with an externally applied magnetic field $B_{0}$ = 9 T. We find that the magnetic dipole field component ($B_{||}^{\text{dip}}$) parallel to $B_{0}$ at the Li site from the Mo electrons has no lattice axial symmetry; it is small around the middle between the lattice $c$ and $a$ axes in the $ac$-plane with the minimum at the field orientation angle $θ$ = +52.5$^{\circ}$, while the $B_{||}^{\text{dip}}$ maximum is at $θ$ = +142.5$^{\circ}$ when $B_{0}$ is applied perpendicular to $b$ ($B_{0}$ $\perp$ $b$), where $θ$ = 0$^{\circ}$ represents the direction of $B_{0}$ $\parallel$ $c$. Further estimate indicates that $B_{||}^{\text{dip}}$ has a maximum value of 0.35 G at $B_{0}$ = 9 T, and the Mo ions have a possible effective magnetic dipole moment 0.015 $μ_{\text{B}}$ per ion, which is significantly smaller than that of a spin 1/2 free electron. By minimizing potential magnetic contributions to the NMR spectrum satellites with the NMR spectroscopy measurements at the direction where the value of the magnetic dipole field is the smallest, the behavior of the independent charge contributions is observed. This work demonstrates that the magnetic dipole field from the Mo electrons is the dominant source of the local magnetic fields at the Li site, and it suggests that the mysterious "metal-insulator" crossover at low temperatures is not a charge effect. The work also reveals valuable local field information for further NMR investigation which is suggested recently [Phys. Rev. B $\bf{85}$, 235128 (2012)] to be key important to the understanding of many mysterious properties of this Q1D material of particular interest.

cond-mat.str-el

Internal static electric and magnetic field at the copper cite in a single crystal of the electron-doped high-T$_{c}$ superconductor Pr$_{1.85}$Ce$_{0.15}$CuO$_{4-y}$

We report $^{63, 65}$Cu-NMR spectroscopy and Knight shift measurements on a single crystal of the electron-doped high-$T_{c}$ superconductor Pr$_{1.85}$Ce$_{0.15}$CuO$_{4-y}$ (PCCO) with an applied magnetic field ($H$) up to 26.42 T. A very small NQR frequency is obtained with the observation of the spectrum, which shows an extremely wide continuous distribution of it that becomes significant narrower below 20 K at $H$ $\parallel$ $c$ where the superconductivity is completely suppressed, indicating a significant change in the charge distribution at the Cu site, while the corresponding changes at $H$ $\perp$ $c$ is negligible when the superconductivity is present or not fully suppressed. The Knight shift and central linewidth are proportional to the applied magnetic field with a high anisotropy. We find that the magnitude of the internal static magnetic field at the copper is dominated by the anisotropic Cu$^{2+}$ 3$d$-orbital contributions, while its weak temperature-dependence is mainly determined by the isotropic contact hyperfine coupling to the paramagnetic Pr$^{3+}$ spins, which also gives rise to the full distribution of the internal static magnetic field at the copper for $H$ $\perp$ $c$. This internal static electric and magnetic field environment at the copper is very different from that in the hole-doped cuprates, and may provide new insight into the understanding of high-$T_{c}$ superconductivity. Other experimental techniques are needed to verify whether the observed significant narrowing of the charge distribution at the Cu site with $H$ $\parallel$ $c$ is caused by the charge ordering (CO) [E. H. da Silva Neto $et ~al.$, to be published in Science] \cite{ehdsn} or a new type of charge modulation.

cond-mat.supr-con

Superconducting anisotropy in the electron-doped high-T$_{c}$ superconductors Pr$_{2-x}$Ce$_{x}$CuO$_{4-y}$

We report superconducting anisotropy measurements in the electron-doped high-$T_{c}$ superconductors (HTSCs) Pr$_{2-x}$Ce$_{x}$CuO$_{4-y}$ (PCCO, $x$ = 0.15 and 0.17) with applied magnetic field ($H_{0}$) up to 28 T. Our results show that the upper critical field [$H_{c2}(T)$] is highly anisotropic, and as temperature $T$ $\rightarrow$ 0 the value of it at $H_{0}$ $\parallel$ $c$ [$H_{c2,\parallel c}$(0)] is far less than the Pauli limit, which is very different from that at $H_{0}$ $\perp$ $c$. The $H_{c2}(0)$ character along with the evaluated zero $T$ coherence length $[ξ_{ab(c)}(0)]$ and penetration depth $[λ_{ab(c)}(0)]$ is compared with those of the hole-doped cuprate HTSCs and typical Fe-based superconductors. We find that the low temperature anisotropic character of PCCO is rather similar to that of hole-doped cuprate HTSCs, but apparently larger than that of typical Fe-based superconductors. This study also proves a new sensitive probe of detecting rich properties of unconventional superconductors with the use of the resonant frequency of a NMR probe circuit.

cond-mat.supr-con

Distribution of Non-uniform Demagnetization Fields in Paramagnetic Bulk Solids

A general calculation for the distribution of non-uniform demagnetization fields in paramagnetic bulk solids is described and the fields for various sample geometries are calculated. Cones, ellipsoids, paraboloids and hyperboloids with similar sample aspect ratios are considered. Significant differences in their demagnetization fields are observed. The calculation shows that the demagnetization field magnitudes decrease along the axis of symmetry (along $z$) where an externally applied magnetic field is aligned, and increase in the vicinity of the lateral surfaces with the largest field values found in the cone and the narrowest field distributions found in the hyperboloid. Application is made to the theoretical modeling of the $^{1}$H-NMR spectra of a single crystal of field-induced superconductor $λ$-(BETS)$_{2}$FeCl$_{4}$ with a rectangular sample geometry, providing a good fit to the measured NMR spectra. This calculation is also applicable to diamagnetic or ferromagnetic materials in general.

cond-mat.str-el

$^{77}$Se NMR measurements of the $π-d$ exchange field in the organic conductor $λ-$(BETS)$_{2}$FeCl$_{4}$

$^{77}$Se-NMR spectrum and frequency shift measurements in the paramagnetic metal (PM) and antiferromagnetic insulating (AFI) phases are reported for a small single crystal of the organic conductor $λ-$(BETS)$_{2}$FeCl$_{4}$ as a function of temperature ($T$) and field alignment for an applied magnetic field $B_{0}$ = 9 T. The results show that in the low $T$ limit, where the localized Fe$^{3+}$ spins ($S_{d}$ = 5/2) are almost fully polarized, the conduction electrons (Se $π$-electrons, spin $s_π$ = 1/2) in the BETS molecules experience an exchange field ($\bf{B}$$_{πd}$) from the Fe$^{3+}$ spins with a value of $-$ 32.7 $\pm$ 1.5 T at 5 K and 9 T aligned opposite to $\bf{B}$$_{0}$. This large negative value of $\bf{B}$$_{πd}$ is consistent with that predicted by the resistivity measurements and supports the Jaccarino-Peter internal field-compensation mechanism being responsible for the origin of field-induced superconductivity.

cond-mat.str-el

$^{1}$H-NMR spin-echo measurements of the static and dynamic spin properties in $λ$-(BETS)$_{2}$FeCl$_{4}$

$^{1}$H-NMR spin-echo measurements of the spin-echo decay $M(2τ)$ with a decay rate 1/$T_{2}$ and the frequency shift $Δν/ν_{0}$ under applied magnetic field $\mathbf{B}$$_{0}$ = 9 T along the a-axis over a temperature range 2.0$-$180 K are reported for a single crystal of the organic conductor $λ$-(BETS)$_{2}$FeCl$_{4}$. It provides the spin dynamic and static properties in the paramagnetic metal (PM) and antiferromagnetic insulator (AFI) states as well as across the PM$-$AFI phase transition. A large slow beat structure in the spin-echo decay is observed with a typical beat frequency of $f$ $\sim$ 7 kHz and it varies across the spectrum. Its origin is attributed to the $^{1}$H$-$$^{1}$H dipole interactions rather than to the much larger dipolar field contribution from the Fe$^{3+}$ electrons (spin $S$ = 5/2). A simple phenomenological model provides an excellent fit to the data. The dominant $^{1}$H-NMR frequency shift comes from the dipolar field from the 3d Fe$^{3+}$ ions, and the Fe$^{3+}$ $-$ Fe$^{3+}$ exchange interactions ($J_{0}$) ($J_{0} $ includes the d$-$d exchange interactions through the $π-$electrons) have a substantial effect to the local field at the proton sites expecially at low temperatures. A good fit is obtained with $J_{0}$ = - 1.7 K. The data of the spin-echo decay rate 1/$T_{2}$ indicates that there is a significant change in the slow fluctuations of the local magnetic field at the $^{1}$H-sites on traversing the PM to AFI phase. This evidence supports earlier reports that the PM$-$AFI phase transition in $λ$-(BETS)$_{2}% $FeCl$_{4} $ is driven magnetically and first order.

cond-mat.str-el

Proton NMR measurements of the local magnetic field in the paramagnetic metal and antiferromagnetic insulator phases of $λ$-(BETS)$_{2}$FeCl$_{4}$

Measurements of the $^{1}$H-NMR spectrum of a small ($\sim$ 4 $μ$g) single crystal of the organic conductor $λ$-(BETS)$_{2}$FeCl$_{4}$ are reported with an applied magnetic field $\bf{B}$$_{0}$ = 9 T parallel to the a-axis in the $ac$-plane over a temperature $(T)$ range 2.0 $-$ 180 K. They provide the distribution of the static local magnetic field at the proton sites in the paramagnetic metal (PM) and antiferromagnetic insulator (AFI) phases, along with the changes that occur at the PM$-$AFI phase transition. The spectra have six main peaks that are significantly broadened and shifted at low $T$. The origin of these features is attributed to the large dipolar field from the 3d Fe$^{3+}$ ion moments (spin $S_{\rm{d}}$ = 5/2). Their amplitude and $T-$dependence are modeled using a modified Brillouin function that includes a mean field approximation for the total exchange interaction ($J_{0}$) between one Fe$^{3+}$ ion and its two nearest neighbors. A good fit is obtained using $J_{0}$ = $-$ 1.7 K. At temperatures below the PM$-$AFI transition temperature $T_{MI}$ = 3.5 K, an extra peak appears on the high frequency side of the spectrum and the details of the spectrum become smeared. Also, the rms linewidth and the frequency shift of the spectral distribution are discontinuous, consistent with the transition being first-order. These measurements verify that the dominant local magnetic field contribution is from the Fe$^{3+}$ ions and indicate that there is a significant change in the static local magnetic field distribution at the proton sites on traversing the PM to AFI phase transition.

cond-mat.str-el

NMR Investigation of the Organic Conductor lambda-(BETS)2FeCl4

The two-dimensional organic conductor lambda-(BETS)2FeCl4 has an unusual phase diagram as a function of temperature and magnetic field that includes a paramagnetic metal (PM) phase, an antiferromagnetic insulating (AFI) phase, and a field-induced superconducting phase [S. Uji, H. Kobayashi, L. Balicas, and James S. Brooks, Adv. Mater. 14, 243 (2002), and cited references]. Here, we report a preliminary investigation of the PM and AFI phases at 9.0 T over the temperature range 2.0-180 K that uses proton NMR measurements of the spectrum, the spin-lattice relaxation rate (1/T1), and the spin echo decay rate (1/T2). The sample is asmall single crystal whose mass is approximately 3 micrograms (approximately 2E16 protons). Its small size creates several challenges that include detecting small signals and excluding parasitic proton signals that are not from the sample [H. N. Bachman and I. F. Silvera, J. Mag. Res. 162, 417 (2003)]. These strategies and other techniques used to obtain viable signals are described.

cond-mat.str-el