A5076697411- AI in cancer detection
- Medical Image Segmentation Techniques
- Radiomics and Machine Learning in Medical Imaging
- Image Retrieval and Classification Techniques
- Cell Image Analysis Techniques
- Advanced Neuroimaging Techniques and Applications
- Advanced Image and Video Retrieval Techniques
- Advanced Neural Network Applications
- Advanced MRI Techniques and Applications
- Image Processing Techniques and Applications
- Medical Imaging Techniques and Applications
- Brain Tumor Detection and Classification
- COVID-19 diagnosis using AI
- Digital Imaging for Blood Diseases
- Domain Adaptation and Few-Shot Learning
- Multimodal Machine Learning Applications
- Functional Brain Connectivity Studies
- Fetal and Pediatric Neurological Disorders
- MRI in cancer diagnosis
- Medical Imaging and Analysis
- 3D Shape Modeling and Analysis
- Biomedical Text Mining and Ontologies
- Lung Cancer Diagnosis and Treatment
- 3D Surveying and Cultural Heritage
- Human Pose and Action Recognition
The University of Sydney
2016-2025
Chongqing University
2025
Chinese Academy of Sciences
2004-2024
Betta Pharmaceuticals (China)
2024
Wenzhou Medical University
2008-2023
Adrian College
2023
Directorate of Medicinal and Aromatic Plants Research
2023
China Iron and Steel Research Institute Group
2023
Nanjing Medical University
2023
Jiangsu Province Hospital
2023
Image patch classification is an important task in many different medical imaging applications. In this work, we have designed a customized Convolutional Neural Networks (CNN) with shallow convolution layer to classify lung image patches interstitial disease (ILD). While feature descriptors been proposed over the past years, they can be quite complicated and domain-specific. Our CNN framework can, on other hand, automatically efficiently learn intrinsic features from that are most suitable...
The accurate diagnosis of Alzheimer's disease (AD) is essential for patient care and will be increasingly important as modifying agents become available, early in the course disease. Although studies have applied machine learning methods computer-aided AD, a bottleneck diagnostic performance was shown previous methods, due to lacking efficient strategies representing neuroimaging biomarkers. In this study, we designed novel framework with deep architecture aid AD. This uses zero-masking...
In this paper, we propose a new clustering model, called DEeP Embedded Regularized ClusTering (DEPICT), which efficiently maps data into discriminative embedding subspace and precisely predicts cluster assignments. DEPICT generally consists of multinomial logistic regression function stacked on top multi-layer convolutional autoencoder. We define objective using relative entropy (KL divergence) minimization, regularized by prior for the frequency An alternating strategy is then derived to...
The accurate diagnosis of Alzheimer's disease (AD) plays a significant role in patient care, especially at the early stage, because consciousness severity and progression risks allows patients to take prevention measures before irreversible brain damages are shaped. Although many studies have applied machine learning methods for computer-aided-diagnosis (CAD) AD recently, bottleneck performance was shown most existing researches, mainly due congenital limitations chosen models. In this...
The accurate identification of malignant lung nodules on chest CT is critical for the early detection cancer, which also offers patients best chance cure. Deep learning methods have recently been successfully introduced to computer vision problems, although substantial challenges remain in due lack large training data sets. In this paper, we propose a multi-view knowledge-based collaborative (MV-KBC) deep model separate from benign using limited data. Our learns 3-D nodule characteristics by...
Discrete point cloud objects lack sufficient shape descriptors of 3D geometries. In this paper, we present a novel method for aggregating hypothetical curves in clouds. Sequences connected points (curves) are initially grouped by taking guided walks the clouds, and then subsequently aggregated back to augment their pointwise features. We provide an effective implementation proposed aggregation strategy including curve grouping operator followed operator. Our was benchmarked on several...
Abstract A heterogeneous cloud system, for example, a Hadoop 2.6.0 platform, provides distributed but cohesive services with rich features on large‐scale management, reliability, and error tolerance. As big data processing is concerned, newly built clusters meet the challenges of performance optimization focusing faster task execution more efficient usage computing resources. Presently proposed approaches concentrate temporal improvement, that is, shortening MapReduce time, seldom focus...
In this paper, we propose a new classification method for five categories of lung tissues in high-resolution computed tomography (HRCT) images, with feature-based image patch approximation. We design two feature descriptors higher descriptiveness, namely the rotation-invariant Gabor-local binary patterns (RGLBP) texture descriptor and multi-coordinate histogram oriented gradients (MCHOG) gradient descriptor. Together intensity features, each is then labeled based on its approximation from...
In recent saliency detection research, many graph-based algorithms have applied boundary priors as background queries, which may generate completely "reversed" maps if the salient objects are on image boundaries. Moreover, these usually depend heavily pre-processed superpixel segmentation, lead to notable degradation in detail features. this paper, a novel method is proposed overcome above issues. First, we propose reversion correction process, locates and removes boundary-adjacent...
Automatic image categorization has become increasingly important with the development of Internet and growth in size databases. Although can be formulated as a typical multi-class classification problem, two major challenges have been raised by real-world images. On one hand, though using more labeled training data may improve prediction performance, obtaining labels is time consuming well biased process. other visual descriptors proposed to describe objects scenes appearing images different...
With the developments of DNA sequencing technology, large amounts data have become available in recent years and provide unprecedented opportunities for advanced association studies between somatic point mutations cancer types/subtypes, which may contribute to more accurate mutation based classification (SMCC). However existing SMCC methods, issues like high sparsity, small volume sample size, application simple linear classifiers, are major obstacles improving performance. To address...
Accurate and reliable segmentation of the prostate gland using magnetic resonance (MR) imaging has critical importance for diagnosis treatment diseases, especially cancer. Although many automated approaches, including those based on deep learning have been proposed, performance still room improvement due to large variability in image appearance, interference, anisotropic spatial resolution. In this paper, we propose 3D adversarial pyramid convolutional neural network (3D APA-Net) MR images....
Channel pruning is a class of powerful methods for model compression. When neural network, it's ideal to obtain sub-network with higher accuracy. However, does not necessarily have high accuracy low classification loss (loss-metric mismatch). In the paper, we first consider loss-metric mismatch problem and propose novel channel method Convolutional Neural Networks (CNNs) by directly maximizing performance (i.e., accuracy) sub-networks. Specifically, train stand-alone network predict...
The task of segmenting cell nuclei and cytoplasm in pap smear images is one the most challenging tasks automated cervix cytological analysis due to specifically presence overlapping cells. This paper introduces a multi-pass fast watershed-based method (MPFW) segment both nucleus from large masses cervical cells three watershed passes. first pass locates with barrier-based on gradient-based edge map pre-processed image. next segments isolated, touching, partially transform adapted shape...
Radiography imaging protocols focus on particular body regions, therefore producing images of great similarity and yielding recurrent anatomical structures across patients. To exploit this structured information, we propose the use Space-aware Memory Queues for In-painting Detecting anomalies from radiography (abbreviated as SQUID). We show that SQUID can taxonomize ingrained into patterns; in inference, it identify (unseen/modified patterns) image. surpasses 13 state-of-the-art methods...