visualization

[timingisnow]

I trying to use this code for real-time stereo 3D object detection.

and I have met problems:

1. First these RTS3D and RTM3D which generate monocular 3D object detection for initial data configuration version low so that my GPU is not working well. So, Do you have any newest version of these code?

2. I tried visualization using below command:
   python ./src/demo.py --vis --demo ./kitti_format/data/kitti/val.txt --data_dir ./kitti_format --calib_dir ./kitti_format/data/kitti/calib/ --load_model ./kitti_format/exp/RTS3D/model_last.pth --gpus 0 --mono_path ./kitti_format/data/kitti/mono_results/

but command is now working (and I tried to find the problems, but I can't find visualization part of executed file(demo.py))

3. anyways, I made visualization code using vis_3d_utils.py for "Project 3D bounding box to rgb frontview for simply visualization" and I found the difference of accurate result with it of paper. your paper, Figure 4 shows more accurate result than monocular. how can I develop the code for more accurate result?

thansk