arXiv subjects

Hao Quan

Publications and source records attributed to Hao Quan.

9 recordsLinked to original sources

arXiv2025

Establishing Causal Relationship Between Whole Slide Image Predictions and Diagnostic Evidence Subregions in Deep Learning

Due to the lack of fine-grained annotation guidance, current Multiple Instance Learning (MIL) struggles to establish a robust causal relationship between Whole Slide Image (WSI) diagnosis and evidence sub-images, just like fully supervised learning. So many noisy images can undermine the network's prediction. The proposed Causal Inference Multiple Instance Learning (CI-MIL), uses out-of-distribution generalization to reduce the recognition confusion of sub-images by MIL network, without requiring pixelwise annotations. Specifically, feature distillation is introduced to roughly identify the feature representation of lesion patches. Then, in the random Fourier feature space, these features are re-weighted to minimize the cross-correlation, effectively correcting the feature distribution deviation. These processes reduce the uncertainty when tracing the prediction results back to patches. Predicted diagnoses are more direct and reliable because the causal relationship between them and diagnostic evidence images is more clearly recognized by the network. Experimental results demonstrate that CI-MIL outperforms state-of-the-art methods, achieving 92.25% accuracy and 95.28% AUC on the Camelyon16 dataset (breast cancer), while 94.29% accuracy and 98.07% AUC on the TCGA-NSCLC dataset (non-small cell lung cancer). Additionally, CI-MIL exhibits superior interpretability, as its selected regions demonstrate high consistency with ground truth annotations, promising more reliable diagnostic assistance for pathologists.

cs.CV↗

arXiv2025

Distilling High Diagnostic Value Patches for Whole Slide Image Classification Using Attention Mechanism

Multiple Instance Learning (MIL) has garnered widespread attention in the field of Whole Slide Image (WSI) classification as it replaces pixel-level manual annotation with diagnostic reports as labels, significantly reducing labor costs. Recent research has shown that bag-level MIL methods often yield better results because they can consider all patches of the WSI as a whole. However, a drawback of such methods is the incorporation of more redundant patches, leading to interference. To extract patches with high diagnostic value while excluding interfering patches to address this issue, we developed an attention-based feature distillation multi-instance learning (AFD-MIL) approach. This approach proposed the exclusion of redundant patches as a preprocessing operation in weakly supervised learning, directly mitigating interference from extensive noise. It also pioneers the use of attention mechanisms to distill features with high diagnostic value, as opposed to the traditional practice of indiscriminately and forcibly integrating all patches. Additionally, we introduced global loss optimization to finely control the feature distillation module. AFD-MIL is orthogonal to many existing MIL methods, leading to consistent performance improvements. This approach has surpassed the current state-of-the-art method, achieving 91.47% ACC (accuracy) and 94.29% AUC (area under the curve) on the Camelyon16 (Camelyon Challenge 2016, breast cancer), while 93.33% ACC and 98.17% AUC on the TCGA-NSCLC (The Cancer Genome Atlas Program: non-small cell lung cancer). Different feature distillation methods were used for the two datasets, tailored to the specific diseases, thereby improving performance and interpretability.

eess.IV↗

arXiv2025

Weak-Strong Uniqueness and Extreme Wall Events at High Reynolds Number

Singular or weak solutions of the incompressible Euler equations have been hypothesized to account for anomalous dissipation at very high Reynolds numbers and, in particular, to explain the d'Alembert paradox of non-vanishing drag. A possible objection to this explanation is the mathematical property called ``weak-strong uniqueness'', which requires that any admissable weak solution of the Euler equations must coincide with the smooth Euler solution for the same initial data. As an application of the Josephson-Anderson relation, we sketch a proof of conditional weak-strong uniqueness for the potential Euler solution of d'Alembert within the class of strong inviscid limits. We suggest that the mild conditions required for weak-strong uniqueness are, in fact, physically violated by violent eruption of very thin boundary layers. We discuss observational signatures of these extreme events and explain how the small length-scales involved could threaten the validity of a hydrodynamic description.

physics.flu-dyn↗

arXiv2025

Weak-Strong Uniqueness and the d'Alembert Paradox

We prove conditional weak-strong uniqueness of the potential Euler solution for external flow around a smooth body in three space dimensions, within the class of viscosity weak solutions with the same initial data. Our sufficient condition is the vanishing of the streamwise component of the skin friction in the inviscid limit, somewhat weaker than the condition of Bardos-Titi in bounded domains. Because global-in-time existence of the smooth potential solution leads back to the d'Alembert paradox, we argue that weak-strong uniqueness is not a valid criterion for "relevant" notions of generalized Euler solution and that our condition is likely to be violated in the inviscid limit. We prove also that the Drivas-Nguyen condition on uniform continuity at the wall of the normal velocity component implies weak-strong uniqueness within the general class of admissible weak Euler solutions in bounded domains.

math.AP↗

arXiv2024

Inertial Momentum Dissipation for Viscosity Solutions of Euler Equations: External Flow Around a Smooth Body

We study the local balance of momentum for weak solutions of incompressible Euler equations obtained from the zero-viscosity limit in the presence of solid boundaries, taking as an example flow around a finite, smooth body. We show that both viscous skin friction and wall pressure exist in the inviscid limit as distributions on the body surface. We define a nonlinear spatial flux of momentum toward the wall for the Euler solution, and show that wall friction and pressure are obtained from this momentum flux in the limit of vanishing distance to the wall, for the wall-parallel and wall-normal components, respectively. We show furthermore that the skin friction describing anomalous momentum transfer to the wall will vanish if velocity and pressure are bounded in a neighborhood of the wall and if also the essential supremum of wall-normal velocity within a small distance of the wall vanishes with this distance (a precise form of the vanishing wall-normal velocity condition). In the latter case, all of the limiting drag arises from pressure forces acting on the body and the pressure at the body surface can be obtained as the limit approaching the wall of the interior pressure for the Euler solution. As one application of this result, we show that Lighthill's theory of vorticity generation at the wall is valid for the Euler solutions obtained in the inviscid limit. Further, in a companion work, we show that the Josephson-Anderson relation for the drag, recently derived for strong Navier-Stokes solutions, is valid for weak Euler solutions obtained in their inviscid limit.

math-ph↗

arXiv2023

Global Contrast Masked Autoencoders Are Powerful Pathological Representation Learners

Based on digital pathology slice scanning technology, artificial intelligence algorithms represented by deep learning have achieved remarkable results in the field of computational pathology. Compared to other medical images, pathology images are more difficult to annotate, and thus, there is an extreme lack of available datasets for conducting supervised learning to train robust deep learning models. In this paper, we propose a self-supervised learning (SSL) model, the global contrast-masked autoencoder (GCMAE), which can train the encoder to have the ability to represent local-global features of pathological images, also significantly improve the performance of transfer learning across data sets. In this study, the ability of the GCMAE to learn migratable representations was demonstrated through extensive experiments using a total of three different disease-specific hematoxylin and eosin (HE)-stained pathology datasets: Camelyon16, NCTCRC and BreakHis. In addition, this study designed an effective automated pathology diagnosis process based on the GCMAE for clinical applications. The source code of this paper is publicly available at https://github.com/StarUniversus/gcmae.

eess.IV↗

arXiv2023

Dual-channel Prototype Network for few-shot Classification of Pathological Images

In pathology, the rarity of certain diseases and the complexity in annotating pathological images significantly hinder the creation of extensive, high-quality datasets. This limitation impedes the progress of deep learning-assisted diagnostic systems in pathology. Consequently, it becomes imperative to devise a technology that can discern new disease categories from a minimal number of annotated examples. Such a technology would substantially advance deep learning models for rare diseases. Addressing this need, we introduce the Dual-channel Prototype Network (DCPN), rooted in the few-shot learning paradigm, to tackle the challenge of classifying pathological images with limited samples. DCPN augments the Pyramid Vision Transformer (PVT) framework for few-shot classification via self-supervised learning and integrates it with convolutional neural networks. This combination forms a dual-channel architecture that extracts multi-scale, highly precise pathological features. The approach enhances the versatility of prototype representations and elevates the efficacy of prototype networks in few-shot pathological image classification tasks. We evaluated DCPN using three publicly available pathological datasets, configuring small-sample classification tasks that mirror varying degrees of clinical scenario domain shifts. Our experimental findings robustly affirm DCPN's superiority in few-shot pathological image classification, particularly in tasks within the same domain, where it achieves the benchmarks of supervised learning.

cs.CV↗

arXiv2022

Onsager Theory of Turbulence, the Josephson-Anderson Relation, and the D'Alembert Paradox

The Josephson-Anderson relation, valid for the incompressible Navier-Stokes solutions which describe flow around a solid body, instantaneously equates the power dissipated by drag to the flux of vorticity across the flow lines of the potential Euler solution considered by d'Alembert. Its derivation involves a decomposition of the velocity field into this background potential-flow field and a solenoidal field corresponding to the rotational wake behind the body, with the flux term describing transfer from the interaction energy between the two fields and into kinetic energy of the rotational flow. We establish the validity of the Josephson-Anderson relation for the weak solutions of the Euler equations obtained in the zero-viscosity limit, with one transfer term due to inviscid vorticity flux and the other due to a viscous skin-friction anomaly. Furthermore, we establish weak forms of the local balance equations for both interaction and rotational energies. We define nonlinear spatial fluxes of these energies and show that the asymptotic flux of interaction energy to the wall equals the anomalous skin-friction term in the Josephson-Anderson relation. However, when the Euler solution satisfies suitably the no-flow-through condition at the wall, then the anomalous term vanishes. In this case, we can show also that the asymptotic flux of rotational energy to the wall must vanish and we obtain in the rotational wake the Onsager-Duchon-Robert relation between viscous dissipation anomaly and inertial dissipation due to scale-cascade. In this way we establish a precise connection between the Josephson-Anderson relation and the Onsager theory of turbulence, and we provide a novel resolution of the d'Alembert paradox.

math-ph↗

arXiv2022

A Deep Reinforcement Learning Framework for Rapid Diagnosis of Whole Slide Pathological Images

The deep neural network is a research hotspot for histopathological image analysis, which can improve the efficiency and accuracy of diagnosis for pathologists or be used for disease screening. The whole slide pathological image can reach one gigapixel and contains abundant tissue feature information, which needs to be divided into a lot of patches in the training and inference stages. This will lead to a long convergence time and large memory consumption. Furthermore, well-annotated data sets are also in short supply in the field of digital pathology. Inspired by the pathologist's clinical diagnosis process, we propose a weakly supervised deep reinforcement learning framework, which can greatly reduce the time required for network inference. We use neural network to construct the search model and decision model of reinforcement learning agent respectively. The search model predicts the next action through the image features of different magnifications in the current field of view, and the decision model is used to return the predicted probability of the current field of view image. In addition, an expert-guided model is constructed by multi-instance learning, which not only provides rewards for search model, but also guides decision model learning by the knowledge distillation method. Experimental results show that our proposed method can achieve fast inference and accurate prediction of whole slide images without any pixel-level annotations.

eess.IV↗