A5066013671- Advanced Neural Network Applications
- Smart Agriculture and AI
- Spectroscopy and Chemometric Analyses
- Identification and Quantification in Food
- Radiomics and Machine Learning in Medical Imaging
- Plant Disease Management Techniques
- Anomaly Detection Techniques and Applications
- Water Quality Monitoring Technologies
- Smart Systems and Machine Learning
- Blockchain Technology Applications and Security
- Power Transformer Diagnostics and Insulation
- Advanced X-ray and CT Imaging
- Digital Imaging for Blood Diseases
- Power Systems Fault Detection
- Machine Fault Diagnosis Techniques
- Stock Market Forecasting Methods
China Agricultural University
2022-2024
Beihang University
2023
This paper aims to address the increasingly severe security threats in financial systems by proposing a novel attack detection model, Finsformer. model integrates advanced Transformer architecture with innovative cluster-attention mechanism, dedicated enhancing accuracy of behavior counter complex and varied strategies. A key innovation Finsformer lies its effective capture information patterns within transaction data. Comparative experiments traditional deep learning models such as RNN,...
This study addresses the problem of maize disease detection in agricultural production, proposing a high-accuracy method based on Attention Generative Adversarial Network (Attention-GAN) and few-shot learning. The introduces an attention mechanism, enabling model to focus more significant parts image, thereby enhancing performance. Concurrently, data augmentation is performed through (GAN) generate training samples, overcoming difficulties Experimental results demonstrate that this surpasses...
With the evolution of modern agriculture and precision farming, efficient accurate detection crop diseases has emerged as a pivotal research focus. In this study, an interpretative high-precision rice disease method, integrating multisource data transfer learning, is introduced. This approach harnesses diverse types, including imagery, climatic conditions, soil attributes, facilitating enriched information extraction enhanced accuracy. The incorporation learning bestows model with robust...
Computed tomography (CT) is the first modern slice-imaging modality. Recent years have witnessed its widespread application and improvement in detecting diagnosing related lesions. Nonetheless, there are several difficulties lesions CT images: (1) image quality degrades as radiation dose reduced to decrease radiational injury human body; (2) frequently hampered by noise interference; (3) because of complicated circumstances diseased tissue, lesion pictures typically show complex shapes; (4)...
With the widespread application of drone technology, demand for pest detection and identification from low-resolution noisy images captured with drones has been steadily increasing. In this study, a lightweight model based on Transformer super-resolution sampling techniques is introduced, aiming to enhance accuracy under challenging conditions. The was found effectively capture spatial dependencies in images, while technique employed restore image details subsequent processes. experimental...
With the development of computer science technology, theory and method image segmentation are widely used in fish discrimination, which plays an important role improving efficiency fisheries sorting biodiversity studying. However, existing methods images less accurate inefficient, is worthy in-depth exploration. Therefore, this paper proposes atrous pyramid GAN network aimed at increasing accuracy efficiency. This introduces structure, module added before CNN backbone order to augment...
In order to improve the transferability of transmission line fault identification models, this paper divides lines into source and target based on transfer learning theory, proposes a method for identifying types deep-transfer learning. First, time series data during faults is generated by combining different conditions, preprocessing data, obtained input samples Convolutional Neural Network (CNN). Second, initial convolutional neural network pre-trained using obtain model type...