A5069751114- Handwritten Text Recognition Techniques
- Vehicle License Plate Recognition
- Image Processing and 3D Reconstruction
- Computational Geometry and Mesh Generation
- Advanced Numerical Analysis Techniques
- Hydrology and Watershed Management Studies
- Forensic Anthropology and Bioarchaeology Studies
- Remote Sensing and LiDAR Applications
- Infrastructure Maintenance and Monitoring
- 3D Surveying and Cultural Heritage
- Digital Media Forensic Detection
- Robotics and Sensor-Based Localization
- Advanced Vision and Imaging
- Advanced Image and Video Retrieval Techniques
- Flood Risk Assessment and Management
- Hydrology and Drought Analysis
- Geophysical Methods and Applications
- Natural Language Processing Techniques
Anyang Normal University
2025
China University of Geosciences
2021-2022
PLA Information Engineering University
2020
Zhengzhou University
2020
Pakistan is a flood-prone country and almost every year, it hit by floods of varying magnitudes. This study was conducted to generate flash flood map using analytical hierarchy process (AHP) frequency ratio (FR) models in the ArcGIS 10.6 environment. Eight flash-flood-causing physical parameters were considered for this study. Five based on digital elevation model (DEM), Advanced Land Observation Satellite (ALOS), Sentinel-2 satellite, including distance from river drainage density slope,...
The goal of this paper is to serve as a guide for selecting detection architecture Oracle Bone Inscription (OBI) that achieves the right speed/memory/accuracy balance given platform. Many successful systems have been proposed in recent years, but directly migrating these methods OBI data may lead unsatisfying performance due corrosion, noise, and distribution. We present unified implementation Faster R-CNN [1], SSD [2], YOLOv3 [3], RFBnet [4], RefineDet [5] which we view experiment...
The detection of Oracle Bone Inscription (OBI) is one the most fundamental aspects oracle bone morphology. However, method depending on experts' experience requires longterm learning and accumulation for professional knowledge. This paper investigated performance deep-learning-based object framework in OBI dataset, then selected with best as baseline made a series optimization. Specifically, we first redesigned sizes ratios anchor box according to data characteristics by using K- means...
Scene text detection methods based on deep learning have recently shown remarkable improvement. Most train convolutional neural networks with full masks requiring pixel accuracy for good quality training. Normally, a skilled engineer needs to drag tens of points create mask the curved text. Therefore, data labelling is time consuming and laborious, particularly texts. To reduce cost, weakly supervised method first proposed in this paper. Unlike other detectors (e.g ., PSENet or TextSnake)...
In this paper, we propose a pixel-wise method named TextCohesion for scene text detection, which splits instance into five key components: Text Skeleton and four Directional Pixel Regions. These components are easier to handle than the entire instance. A confidence scoring mechanism is designed filter characters that similar text. Our can integrate contexts intensively when backgrounds complex. Experiments on two curved challenging benchmarks demonstrate outperforms state-of-the-art methods,...
Floor plan vectorization is an emerging research area in geographic information science and computer vision. However, automated recognition of building elements remains a challenge. This work proposes method that combines the advantages classical graphics with deep learning. Specifically, morphological template introduced to optimize topological relations, enhance completeness, suppress conflicts. Bezier curves are utilized represent irregularity contributing improving visual effects...
The performance of text detection is crucial for the subsequent recognition task. Currently, accuracy detector still needs further improvement, particularly those with irregular shapes in a complex environment. We propose pixel-wise method based on instance segmentation scene detection. Specifically, split into five components: Text Skeleton and four Directional Pixel Regions, then restoring itself these elements receiving supplementary information from other areas when one fails. Besides,...