A5040555643- Medical Image Segmentation Techniques
- Image Retrieval and Classification Techniques
- Advanced Image Fusion Techniques
- Colorectal Cancer Screening and Detection
- Metabolomics and Mass Spectrometry Studies
- AI in cancer detection
- Seaweed-derived Bioactive Compounds
- Face and Expression Recognition
- Phytochemistry and Biological Activities
Shenyang University
2024
Introduction “Sweating,” a key step in the processing and production of Eucommiae Cortex (EC), which plays vital role formation medicinal quality EC. However, mechanism effect this traditional treatment herbs on is still unclear. Methods In study, high performance liquid chromatography (HPLC), UPLC/MS-based untargeted metabolomics high-throughput sequencing were applied to investigate dynamic changes main active ingredients, differential metabolites bacterial communities process “sweating”...
Aiming at a series of limitations the Adam algorithm, such as hyperparameter sensitivity and unstable convergence, in this paper, an improved optimization Cycle-Norm-Adam (CN-Adam) is proposed. The algorithm integrates ideas cyclic exponential decay learning rate (CEDLR) gradient paradigm constraintsand accelerates convergence speed model improves its generalization performance by dynamically adjusting rate. In order to verify effectiveness CN-Adam we conducted extensive experimental...
Due to the complexity and illegibility of medical images, it brings inconvenience difficulty diagnosis personnel. To address these issues, an optimization algorithm called GSL(Gradient sine linear) based on Adam improvement is proposed in this paper, which introduces gradient pruning strategy, periodic adjustment learning rate, linear interpolation strategy. The trimming technique can scale prevent explosion, while rate strategy adjusts according characteristics sinusoidal function,...
In this study, a gastrointestinal image classification method based on the improved Adam algorithm is proposed. Gastrointestinal of great significance in field medical analysis, but it presents numerous challenges, including slow convergence, susceptibility to local minima, and complexity imbalance data. Although widely used stochastic gradient descent, tends suffer from overfitting explosion issues when dealing with complex To address these problems, paper proposes an algorithm, AdamW_AGC,...