Feature-based Visual Odometry for Bronchoscopy: A Dataset and Benchmark [IROS'23]

---

Installation of PySLAM

Please check this file for more detail instructions.

---

Bronchoscopy dataset

Here's the link to download our dataset: link

---

Usage for bronchoscopy dataset

1. Link the data folder to ./dataset in this architecture

   PYSLAM
   |__ main_vo.py
   |__ config.ini
   |__ ....
   |___ dataset
       |__ bronchoscopy
           |__ real_seq_xxx_part_x_dif_x.mp4
           |__ real_seq_xxx_part_x_dif_x.txt
           |__ ....
           
   

2. Change your config.ini file accordingly. Find the section for [FINAL_DATASET], change the base path to your project folder. For example: base_path = ~/PYSLAM/dataset. Then change the tag=bronchoscopy. Select the file for evaluation by setting the keyword fname=file_to_evaluation. Choose the correcting camera setting file in cam_settings=settings/bronchoscopy.yaml. While kept other keyword unchanged.

3. Dependencies: mannually install those libraries if you have module not found error.

4. Camera parameters is stored in ./settings/*.yaml, make sure you are using the right camera parameter file

5. Figures will be shown when you run the program. Mannualy save those you think are useful. Press Q on the image window to close those figures. Evaluation for APE will be shown after closing the window, save them for further evaluation.

6. Further information about evaluation, check the repo evo on github.

---