A5044307829- AI in cancer detection
- Digital Imaging for Blood Diseases
- Radiomics and Machine Learning in Medical Imaging
- Image Processing Techniques and Applications
- Cell Image Analysis Techniques
- Multimodal Machine Learning Applications
- Medical Image Segmentation Techniques
- Advanced Image and Video Retrieval Techniques
- Domain Adaptation and Few-Shot Learning
- Advanced Neural Network Applications
- Biomedical Text Mining and Ontologies
- Topic Modeling
- Cerebrovascular and Carotid Artery Diseases
- Human Pose and Action Recognition
- Genomic variations and chromosomal abnormalities
- COVID-19 diagnosis using AI
- Machine Learning and Data Classification
- Genomics and Chromatin Dynamics
- Advanced Photocatalysis Techniques
- Image Retrieval and Classification Techniques
- Retinal Imaging and Analysis
- Image and Object Detection Techniques
- Cervical Cancer and HPV Research
- Optical measurement and interference techniques
- Data Mining Algorithms and Applications
Westlake University
2021-2025
Institute for the Future
2023
Institute for Advanced Study
2022
Zhejiang University of Technology
2018-2022
Zhejiang University of Science and Technology
2022
Nanjing Tech University
2020
Nanyang Technological University
2020
As advances in large language models (LLMs) and multimodal techniques continue to mature, the development of general-purpose (MLLMs) has surged, offering significant applications interpreting natural images. However, field pathology largely remained untapped, particularly gathering high-quality data designing comprehensive model frameworks. To bridge gap MLLMs, we present PathAsst, a generative foundation AI assistant revolutionize diagnostic predictive analytics pathology. The PathAsst...
While Multiple Instance Learning (MIL) has shown promising results in digital Pathology Whole Slide Image (WSI) analysis, such a paradigm still faces performance and generalization problems due to high computational costs limited supervision of Gigapixel WSIs. To deal with the computation problem, previous methods utilize frozen model pretrained from ImageNet obtain representations, however, it may lose key information owing large domain gap hinder ability without image-level training-time...
Based on many studies, trichosanthin (TCS) has an antiviral effect that regulates immune response, and targets cancer cells to exert broad-spectrum anti-tumor pharmacological activities. It is speculated TCS may be a potential natural active drug for preventing as well treating cervical cancer. But the clearer impact along with underlying mechanism are still unclear. The purpose of this study investigate function in We measured viability cell lines (HeLa & caski cells) using CCK-8 analysis,...
The photocatalytic performance of donor–acceptor structured g-C<sub>3</sub>N<sub>4</sub> was enhanced by up to 3.83-fold due the accelerated intramolecular charge transfer.
Synthetic data generation emerges as a strategy to mitigate scarcity in digital pathology, where complicated tissue and cellular features are correlated with cancer diagnosis. The synthesis of such visuals, however, suffers from limited inter class diversity annotations. Current methodologies struggle capturing the broad spectrum pathology features, causing unpredictable objects defected fidelity. Moreover, discrepancies image resolution across developmental operational phases can amplify...
Abstract:
The emergence of large multimodal models has unlocked remarkable potential in AI, particularly pathology. However, the lack specialized, high-quality benchmark impeded their development and precise evaluation. To address this, we introduce PathMMU, largest highest-quality expert-validated pathology for LMMs. It comprises 33,573 multi-choice questions 21,599 images from various sources, an explanation correct answer accompanies each question. construction PathMMU capitalizes on robust...
Point-based cell detection (PCD), which pursues high-performance sensing under low-cost data annotation, has garnered increased attention in computational pathology community. Unlike mainstream PCD methods that rely on intermediate density map representations, the Point-to-Point network (P2PNet) recently emerged as an end-to-end solution for PCD, demonstrating impressive accuracy and efficiency. Nevertheless, P2PNet is limited to decoding from a single-level feature due scale-agnostic...
Abstract Accurate and automatic carotid artery segmentation for magnetic resonance (MR) images is eagerly expected, which can greatly assist a comprehensive study of atherosclerosis accelerate the translation. Although many efforts have been made, identification inner lumen outer wall in diseased vessels still challenging task due to complex vascular deformation, blurred boundary, confusing componential expression. In this paper, we introduce novel fully 3D framework simultaneously...
The success of supervised deep learning models in medical image segmentation relies on detailed annotations. However, labor-intensive manual labeling is costly and inefficient, especially dense object segmentation. To this end, we propose a self-supervised based approach with Prior Self-activation Module (PSM) that generates self-activation maps from the input images to avoid costs further produce pseudo masks for downstream task. be specific, firstly train neural network using utilize...
With the advancement of deep learning, computer-assisted clinical diagnosis, such as liquid-based cervical cytology, has attracted more attention. However, fragile robustness learning models a non-negligible impact on their classification accuracy and reliability. To be specific, various scanner parameters will used depending pathologist's preferences during diagnosis process (e.g., field source brightness, contrast, saturation, etc.), this variation lead to unstable performance model. In...
As advances in large language models (LLMs) and multimodal techniques continue to mature, the development of general-purpose (MLLMs) has surged, offering significant applications interpreting natural images. However, field pathology largely remained untapped, particularly gathering high-quality data designing comprehensive model frameworks. To bridge gap MLLMs, we present PathAsst, a generative foundation AI assistant revolutionize diagnostic predictive analytics pathology. The PathAsst...
Whole slide images are the foundation of digital pathology for diagnosis and treatment carcinomas. Writing reports is laborious error-prone inexperienced pathologists. To reduce workload improve clinical automation, we investigate how to generate given whole images. On data end, curated largest WSI-text dataset (TCGA-PathoText). In specific, collected nearly 10000 high-quality pairs visual-language models by recognizing cleaning which narrate diagnostic slides in TCGA. model propose multiple...
Cell classification and counting in immunohistochemical cytoplasm staining images play a pivotal role cancer diagnosis. Weakly supervised learning is potential method to deal with labor-intensive labeling. However, the inconstant cell morphology subtle differences between classes also bring challenges. To this end, we present novel recognition framework based on multi-task learning, which utilizes two additional auxiliary tasks guide robust representation of main task. misclassification,...