A5016430669- AI in cancer detection
- Advanced Neural Network Applications
- Medical Image Segmentation Techniques
- Brain Tumor Detection and Classification
- Radiomics and Machine Learning in Medical Imaging
- Digital Imaging for Blood Diseases
- Forensic Anthropology and Bioarchaeology Studies
- Islamic Finance and Banking Studies
- Infrastructure Maintenance and Monitoring
- Domain Adaptation and Few-Shot Learning
- Image Enhancement Techniques
- Machine Learning and Data Classification
- Imbalanced Data Classification Techniques
- Cutaneous Melanoma Detection and Management
- Insurance and Financial Risk Management
- Image Processing Techniques and Applications
- Advanced Image Processing Techniques
- Risk Management in Financial Firms
- Vehicle License Plate Recognition
- Advanced Image Fusion Techniques
- Advanced Image and Video Retrieval Techniques
- Dental Radiography and Imaging
- Remote-Sensing Image Classification
- Medical Imaging and Analysis
- COVID-19 diagnosis using AI
Chengdu University of Technology
2024
University of Electronic Science and Technology of China
2019-2023
Semi-supervised learning is a machine approach that tackles the challenge of having large set unlabeled data and few labeled ones. In this paper we adopt semi-supervised self-training method to increase amount training data, prevent overfitting improve performance deep models by proposing novel selection algorithm prevents mistake reinforcement which common thing in conventional models. The model leverages, specifically, after each training, first generate pseudo-labels on be added samples....
Traffic sign recognition is a classification problem that poses challenges for computer vision and machine learning algorithms. Although both techniques have constantly been improved to solve this problem, the sudden rise in number of unlabeled traffic signs has become even more challenging. Large data collation labeling are tedious expensive tasks demand much time, expert knowledge, fiscal resources satisfy hunger deep neural networks. Aside from that, having unbalanced also greater...
In hyperspectral image (HSI) classification, Convolutional Neural Networks (CNNs) have exhibited exceptional performance, owing to their hierarchical nonlinear modeling. However, fixed square receptive field constrains ability effectively handle irregular regions. Graph Convolution (GCNs) been introduced learn regions through correlations between adjacent pixels modeled as superpixel-based nodes, yet they lack pixel-level information. We propose a novel approach "Pixel-level with Covariance...
The unavailability of large amounts well-labeled data poses a significant challenge in many medical imaging tasks. Even the likelihood having access to sufficient data, process accurately labeling is an arduous and time-consuming one, requiring expertise skills. Again, issue unbalanced further compounds abovementioned problems presents considerable for machine learning algorithms. In lieu this, ability develop algorithms that can exploit unlabeled together with small amount labeled while...
The study investigates the influence of risk management on organizational efficiency. research was done in Ghana, particular Access Bank Ghana Ltd (UPSA Branch) as case study. objectives were aimed at understanding management, knowing potency identification, assessment and analysis, monitoring controlling with it impact Simple random sampling used to select fifteen respondents. instruments majorly included a set questionnaires for findings multiple linear regression show that an element...
Deep learning methods have widely been applied to accurately identify and segment individual teeth in panoramic X-ray radiographs. However, the task becomes challenging as deep models grow deeper wider. Contextual information pass through many layers leading features vanishing before reaching end of model. This study proposes a learning-based three-step model from radiographs address issues above. Firstly, an automatic binarized transformation images deal with computational complexity...
Automatically classifying skin lesion is a challenging task owing to reasons such as high intra class variations, similarities between inter-class images, occlusions in dermoscopy images that impede accurate localization, not mention data unavailability. Considering unlabeled abundant and cheap, this work proposes classification framework integrates the semi-supervised learning concepts of self-training self-paced lesions. First, instance segmentation performed using Mask R-CNN model...
Liver segmentation is a challenging problem where fully supervised deep learning models require large amounts of voxel-wise labels, which are usually laborious, expensive, and time consuming to obtain. However, massive non-annotated 3D CT volumes easily accessible. Most based works address semi-supervised medical image problems have mostly been graph-based or adversarial-based. In sharp contrast, this work investigates the potential method; novel self-training framework utilizing self-paced...