Junbin He

ORCID: 0009-0005-0301-4162
Publications
Citations
Views
---
Saved
---
About
Contact & Profiles
Research Areas
  • Image Retrieval and Classification Techniques
  • Image Processing Techniques and Applications
  • Digital Imaging for Blood Diseases
  • Gear and Bearing Dynamics Analysis
  • Machine Fault Diagnosis Techniques
  • Remote-Sensing Image Classification
  • Structural Integrity and Reliability Analysis
  • Image Enhancement Techniques
  • Color Science and Applications

Hunan Normal University
2019-2020

Froth color can be referred to as a direct and instant indicator the key flotation production index, for example, concentrate grade. However, it is intractable measure froth robustly due adverse interference of time-varying uncontrollable multisource illuminations in process monitoring. In this article, we proposed an illumination-invariant measuring method by solving structure-preserved image-to-image translation task via introduced Wasserstein distance-based structure-preserving CycleGAN,...

10.1109/tcyb.2020.2977537 article EN IEEE Transactions on Cybernetics 2020-03-18

For the problem of class-imbalance in operation monitoring data wind turbine (WT) pitch connecting bolts, an improved Borderline-SMOTE oversampling method based on “two-step decision” with adaptive selection synthetic instances (TSDAS-SMOTE) is proposed. Then, TSDAS-SMOTE combined XGBoost to construct a WT connection bolt fault detection model. generates new samples by decision making” avoid class–class boundary blurring that tends cause when oversampling. First, nearest neighbor sample...

10.1142/s0218348x23401473 article EN cc-by-nc-nd Fractals 2023-01-01

Edge-relevant structure features (ERSFs), e.g., object edges, boundaries and contours, junctions, etc. play an important role in the low middle level image processing task, such as segmentation, well higher-level computer vision tasks, scene analysis understanding. Commonly-used ERSF detection methods employ integer-order differentiation-based methods, which are noise-sensitive have less selectivity of edge feature. Hence, they difficult to effectively extract especially natural images with...

10.1109/ccdc.2019.8832964 article EN 2019-06-01
Coming Soon ...