Mesh Extraction From mip-splatting output (從mip-splatting的輸出提取網格)

[judy8889107]

你好，謝謝你們提供這麼棒的專案！
Hello, thank you for providing such a great project!

我有個問題:
I have a question:

我使用mip-splatting，訓練並輸出nerf_synthetic的8個場景，輸出的格式跟README文件提及的相同
I used mip-splatting to train and output the 8 scenes of nerf_synthetic. The output format is the same as mentioned in the README file.

- output_dir
    - cameras.json (necessary)
    - point_cloud 
        - iteration_7000
            - point_cloud.ply (necessary)
        - iteration_30000
            - point_cloud.ply (necessary)

我有先確定訓練的模型在mip-splatting的OlineViewer上面檢視是正常的。
I have confirmed that the trained model displays correctly on mip-splatting's OlineViewer.

我使用下列指令提取網格(iter 30000):
I used the following command to extract the mesh (iter 30000):

gs-extract-mesh -m "C:\Users\ESA\Desktop\NTPU\mip-splatting\benchmark_nerf_synthetic_ours_stmt\chair" -o ./mesh_output

會輸出gs_mesh.ply到指定目錄，但結果似乎不理想，請問是我哪個環節有做錯嗎？或者還需要調整哪些參數呢？
It outputs gs_mesh.ply to the specified directory, but the result doesn't seem ideal. Could you help me identify if I made any mistakes in the process, or if there are other parameters I need to adjust?

(下方模型使用MeshLab開啟)
(The model below is opened using MeshLab.)

下方的連結是我輸入和輸出的資料，謝謝您!😊
The link below contains the input and output data I used. Thank you!😊
https://nationaltaipeiuniversity-my.sharepoint.com/:u:/g/personal/s711283112_ms_ntpu_edu_tw/EdN7uq0jbpRIp_FFpLwamHEBqojBiByC-YozZMXHWEPlOg?e=3LRlr3