Project
These are my projects I am developing.
## RDS-SLAM
### Demo
- Real Environment: [Youtube-Mask R-CNN + Kinect](https://youtu.be/-peAZEO6Bbo)
- TUM Dataset (Mask R-CNN): [Youtube-walk xyz](https://youtu.be/P-dew4M5Un0)
- TUM Dataset (SegNet): [Youtube-walk xyz](https://youtu.be/7E7Y5ER11B8)
RDS-SLAM: TUM walk_xyz Dataset Evaluation (Mask R-CNN)
RDS-SLAM: Realtime Demo Using Kinect (Mask R-CNN)
### Paper
Y. Liu and J. Miura, "RDS-SLAM: Real-time Dynamic SLAM using Semantic Segmentation Methods," in IEEE Access, doi: 10.1109/ACCESS.2021.3050617.
```sh
@ARTICLE{9318990,
author={Y. {Liu} and J. {Miura}},
journal={IEEE Access},
title={RDS-SLAM: Real-time Dynamic SLAM using Semantic Segmentation Methods},
year={2021},
volume={},
number={},
pages={1-1},
doi={10.1109/ACCESS.2021.3050617}}
```sh
## RTS-vSLAM
This project aims to do semantic segmentation in real-time using ROS action or service.
There are two modules: client and Server. The client will send the source image to request the semantic label from Server. Client-side does not need to deploy GPU. The server is designed to conduct semantic segmentation using GPU.
客户端: 定閱圖像topic,向服務器端實時發送語義分割請求;
服務端: 對圖像進行segmantic segmentating並返回分割結果;
客戶端與服務器可以運行在不同的主機上，本程序使用ROS進行通信;這樣客戶端並不需要配置GPU.
## MaskRCNN-ROS
- Demo 1
- Demo 2
## Geometry Visualization
Calculate and visualize Quaternion and Rotation matrix using Euler Angle (yaw, pitch, roll).
- Github: https://github.com/yubaoliu/VisualizeGeometry.git
