Running evaluating code generated an error

[00Haocheng]

Thank you very much for sharing this excellent work. I made an error after running it, which was specifically expressed as 3d could not be evaluated. The output is as follows：

=========================== Evaluating room0

Successfully evaluated 3D reconstruction.

Traceback (most recent call last): File "/data/whc/Point-SLAM-main/src/Mapper.py", line 924, in run result_dict = literal_eval(result_dict) File "/home/zju231/anaconda3/envs/point-slam/lib/python3.10/ast.py", line 62, in literal_eval node_or_string = parse(node_or_string.lstrip(" \t"), mode='eval') File "/home/zju231/anaconda3/envs/point-slam/lib/python3.10/ast.py", line 50, in parse return compile(source, filename, mode, flags, File "", line 0

SyntaxError: invalid syntax Failed to evaluate 3D reconstruction.

How can I fix it

[eriksandstroem]

Hi, Sorry for my late reply. Can you import the evaluation package? i.e. from evaluate_3d_reconstruction import run_evaluation?

[eriksandstroem]

Any updates here?

[00Haocheng]

Thank you very much for your reply. I have tried and can import this package

[00Haocheng]

Hello, thank you for your guidance; when I run the following code alone in the terminal, python src/tools/eval_recon.py --rec_mesh --gt_mesh -3d it seems that the evaluation algorithm can run normally; I set gt_mesh to room0.ply and rec_mesh to room0_pred_mesh.ply under the mesh file in output, and it works Evaluated normally;

But I found that the recall and F1 indicators are extremely low, and the precision is extremely high; is this due to the very sparse grid in the room0_pred_mesh.ply file, but I only found this one ply file in the output

[eriksandstroem]

Hi, This is indeed a bit odd. Could it be that you are not running on all frames somehow and that this is why the recall is so low? This should not happen though so would be surprised if this is the reason. Since the recall is low, if you look at the reconstructed mesh, can you see that it is not complete?

[00Haocheng]

I opened the file and looked at it, it should be incomplete and indeed very sparse

I thought this was normal output originally, but I ran the code again, and this time the reconstruction results and evaluation indicators were normal and no longer very sparse; I thought maybe I set some wrong configuration when I ran it for the first time or an accidental bug, no matter what, just rebuild it normally, thank you very much!

[eriksandstroem]

Ah ok ! Thanks for sharing. If you find out what made this happen, let me know!