DEMO_REFINE.md

1. Before you start, please follow the instructions to prepare the dataset as described here.

2. Download the model checkpoint and coarse proposals here. The files in the zipped file have the following structure. ${V-A} denotes the unzipped file directory.

   ${V-A}
   ├── epoch_80.tar (The model checkpoint)
   ├── save_config.py (The experiment configuration file)
   ├── IDA-3D (predictions generated by IDA-3D)
      ├── xxxxxx.txt (IDA-3D predictions for image xxxxxx)
   ├── RTS-3D (predictions generated by RTS-3D)
      ├── xxxxxx.txt (RTS-3D predictions for image xxxxxx)      
   

3. Change directory to tools and run

 python inference_agnostic.py --loadmodel ${V-A}/epoch_80.tar --pred_dir ${V-A}/IDA-3D --devices "0" --debug

This command loads a IDA-3D prediction, refines its pose and visualize the process. You should be able to see figures like below and interact with your mouse.

In this figure, the 2D plots shows the left/right inputs and the projected region-of-interest (Red). The 3D plot shows the current proposal (Red), the region-of-interest (yellow) and the down-sampled voxels (blue).

This 2D figure shows the confidence maps for each object part and the predicted coordinates for part update.

This 3D figure shows the input proposal (red), the region-of-interest (yellow), the ground truth (black) and the prediction of SNVC (magenta). Note the refined center is closer to the ground truth center than the input one.