Init Texture(Init Image) looks weired

[sunwonlikeyou]

I ran the all pre-processing codes. And finally running the {sh_file_name}, but 3d result looks weired. Especially 0-4 views textures are white(or none). and below result is generate_image = init_image.convert("RGBA") from generate_texture.py, line 295, view idx==0. I think initial texture has to have textures from input image. Is this weird? or right?

[admg67]

I ran the all pre-processing codes. And finally running the {sh_file_name}, but 3d result looks weired. Especially 0-4 views textures are white(or none). and below result is generate_image = init_image.convert("RGBA") from generate_texture.py, line 295, view idx==0. I think initial texture has to have textures from input image. Is this weird? or right?

Hello, the following problems occurred during the running of my code, can you help me solve them? (ultraman) root@I1c915cb88800301a87:/hy-tmp/Ultraman# ./bash/run_mesh.sh /hy-tmp/Ultraman/Ult 008 ./result : invalid optionsh: line 1: set: - set: usage: set [-abefhkmnptuvxBCHP] [-o option-name] [--] [arg ...] ./bash/run_mesh.sh: line 2: $'\r': command not found ./bash/run_mesh.sh: line 5: $'\r': command not found ./bash/run_mesh.sh: line 6: $'\r': command not found Start Loading Model ... Model Loaded !! 0%| | 0/1 [00:00<?, ?it/s]2.86102294921875e-06 /usr/local/miniconda3/envs/ultraman/lib/python3.8/site-packages/torch/nn/functional.py:4277: UserWarning: Default grid_sample and affine_grid behavior has changed to align_corners=False since 1.3.0. Please specify align_corners=True if the old behavior is desired. See the documentation of grid_sample for details. warnings.warn( /usr/local/miniconda3/envs/ultraman/lib/python3.8/site-packages/torch/nn/functional.py:4215: UserWarning: Default grid_sample and affine_grid behavior has changed to align_corners=False since 1.3.0. Please specify align_corners=True if the old behavior is desired. See the documentation of grid_sample for details. warnings.warn( 3.6102426052093506 100%|██████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 1/1 [00:15<00:00, 15.56s/it] PyMeshLab 0.1.7 based on MeshLab 2020.12d 0%| | 0/1 [00:00<?, ?it/s]008.ply 0%| | 0/1 [00:00<?, ?it/s] Traceback (most recent call last): File "models/2K2K/test_03_poisson.py", line 57, in main(args.save_path, args.save_name) File "models/2K2K/test_03_poisson.py", line 34, in main ms = pymeshlab.Mesh(vertex_matrix = xyz, v_color_matrix=rgba) TypeError: init(): incompatible constructor arguments. The following argument types are supported:

1. pymeshlab.pmeshlab.Mesh(vertex_matrix: numpy.ndarray[numpy.float32[m, 3]] = array([], shape=(0, 3), dtype=float32), face_matrix: numpy.ndarray[numpy.int32[m, 3]] = array([], shape=(0, 3), dtype=int32), v_normals_matrix: numpy.ndarray[numpy.float32[m, 3]] = array([], shape=(0, 3), dtype=float32), f_normals_matrix: numpy.ndarray[numpy.float32[m, 3]] = array([], shape=(0, 3), dtype=float32))

Invoked with: kwargs: vertex_matrix=TrackedArray([[ 0.0116172 , 0.35717618, 0.03688662], [ 0.01200351, 0.3561042 , 0.0396708 ], [ 0.01240174, 0.35544643, 0.04137919], ..., [-0.06300603, -0.34894037, 0.04522913], [-0.06259257, -0.34896928, 0.04515301], [-0.06218336, -0.34902242, 0.04501301]]), v_color_matrix=TrackedArray([[1., 1., 1., 1.], [1., 1., 1., 1.], [1., 1., 1., 1.], ..., [1., 1., 1., 1.], [1., 1., 1., 1.], [1., 1., 1., 1.]])

[sunwonlikeyou]

I ran the all pre-processing codes. And finally running the {sh_file_name}, but 3d result looks weired. Especially 0-4 views textures are white(or none). and below result is generate_image = init_image.convert("RGBA") from generate_texture.py, line 295, view idx==0. I think initial texture has to have textures from input image. Is this weird? or right?

Hello, the following problems occurred during the running of my code, can you help me solve them? (ultraman) root@I1c915cb88800301a87:/hy-tmp/Ultraman# ./bash/run_mesh.sh /hy-tmp/Ultraman/Ult 008 ./result : invalid optionsh: line 1: set: - set: usage: set [-abefhkmnptuvxBCHP] [-o option-name] [--] [arg ...] ./bash/run_mesh.sh: line 2: $'\r': command not found ./bash/run_mesh.sh: line 5: $'\r': command not found ./bash/run_mesh.sh: line 6: $'\r': command not found Start Loading Model ... Model Loaded !! 0%| | 0/1 [00:00<?, ?it/s]2.86102294921875e-06 /usr/local/miniconda3/envs/ultraman/lib/python3.8/site-packages/torch/nn/functional.py:4277: UserWarning: Default grid_sample and affine_grid behavior has changed to align_corners=False since 1.3.0. Please specify align_corners=True if the old behavior is desired. See the documentation of grid_sample for details. warnings.warn( /usr/local/miniconda3/envs/ultraman/lib/python3.8/site-packages/torch/nn/functional.py:4215: UserWarning: Default grid_sample and affine_grid behavior has changed to align_corners=False since 1.3.0. Please specify align_corners=True if the old behavior is desired. See the documentation of grid_sample for details. warnings.warn( 3.6102426052093506 100%|██████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 1/1 [00:15<00:00, 15.56s/it] PyMeshLab 0.1.7 based on MeshLab 2020.12d 0%| | 0/1 [00:00<?, ?it/s]008.ply 0%| | 0/1 [00:00<?, ?it/s] Traceback (most recent call last): File "models/2K2K/test_03_poisson.py", line 57, in main(args.save_path, args.save_name) File "models/2K2K/test_03_poisson.py", line 34, in main ms = pymeshlab.Mesh(vertex_matrix = xyz, v_color_matrix=rgba) TypeError: init(): incompatible constructor arguments. The following argument types are supported: 1. pymeshlab.pmeshlab.Mesh(vertex_matrix: numpy.ndarray[numpy.float32[m, 3]] = array([], shape=(0, 3), dtype=float32), face_matrix: numpy.ndarray[numpy.int32[m, 3]] = array([], shape=(0, 3), dtype=int32), v_normals_matrix: numpy.ndarray[numpy.float32[m, 3]] = array([], shape=(0, 3), dtype=float32), f_normals_matrix: numpy.ndarray[numpy.float32[m, 3]] = array([], shape=(0, 3), dtype=float32))

Invoked with: kwargs: vertex_matrix=TrackedArray([[ 0.0116172 , 0.35717618, 0.03688662], [ 0.01200351, 0.3561042 , 0.0396708 ], [ 0.01240174, 0.35544643, 0.04137919], ..., [-0.06300603, -0.34894037, 0.04522913], [-0.06259257, -0.34896928, 0.04515301], [-0.06218336, -0.34902242, 0.04501301]]), v_color_matrix=TrackedArray([[1., 1., 1., 1.], [1., 1., 1., 1.], [1., 1., 1., 1.], ..., [1., 1., 1., 1.], [1., 1., 1., 1.], [1., 1., 1., 1.]])

I guess you didn't finish 'test_02_model.py' well . It looks empty mesh.

[tomorrow1238]

I ran the all pre-processing codes. And finally running the {sh_file_name}, but 3d result looks weired. Especially 0-4 views textures are white(or none). and below result is generate_image = init_image.convert("RGBA") from generate_texture.py, line 295, view idx==0. I think initial texture has to have textures from input image. Is this weird? or right?

Thank you very much for supporting our work. You can check if the 0-4 views images are generated properly. We will be working on optimizing our pipeline in the near future. we will update a version of the code when we are done with the changes.

[tomorrow1238]

I ran the all pre-processing codes. And finally running the {sh_file_name}, but 3d result looks weired. Especially 0-4 views textures are white(or none). and below result is generate_image = init_image.convert("RGBA") from generate_texture.py, line 295, view idx==0. I think initial texture has to have textures from input image. Is this weird? or right?

Hello, the following problems occurred during the running of my code, can you help me solve them? (ultraman) root@I1c915cb88800301a87:/hy-tmp/Ultraman# ./bash/run_mesh.sh /hy-tmp/Ultraman/Ult 008 ./result : invalid optionsh: line 1: set: - set: usage: set [-abefhkmnptuvxBCHP] [-o option-name] [--] [arg ...] ./bash/run_mesh.sh: line 2:

′

\r
′

: c o m m a n d n o t f o u n d .

/

b a s h

/

r u

n
m

e s h . s h : l i n e 5 : '\r': command not found ./bash/run_mesh.sh: line 6: $'\r': command not found Start Loading Model ... Model Loaded !! 0%| | 0/1 [00:00<?, ?it/s]2.86102294921875e-06 /usr/local/miniconda3/envs/ultraman/lib/python3.8/site-packages/torch/nn/functional.py:4277: UserWarning: Default grid_sample and affine_grid behavior has changed to align_corners=False since 1.3.0. Please specify align_corners=True if the old behavior is desired. See the documentation of grid_sample for details. warnings.warn( /usr/local/miniconda3/envs/ultraman/lib/python3.8/site-packages/torch/nn/functional.py:4215: UserWarning: Default grid_sample and affine_grid behavior has changed to align_corners=False since 1.3.0. Please specify align_corners=True if the old behavior is desired. See the documentation of grid_sample for details. warnings.warn( 3.6102426052093506 100%|██████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 1/1 [00:15<00:00, 15.56s/it] PyMeshLab 0.1.7 based on MeshLab 2020.12d 0%| | 0/1 [00:00<?, ?it/s]008.ply 0%| | 0/1 [00:00<?, ?it/s] Traceback (most recent call last): File "models/2K2K/test_03_poisson.py", line 57, in main(args.save_path, args.save_name) File "models/2K2K/test_03_poisson.py", line 34, in main ms = pymeshlab.Mesh(vertex_matrix = xyz, v_color_matrix=rgba) TypeError: init(): incompatible constructor arguments. The following argument types are supported: 1. pymeshlab.pmeshlab.Mesh(vertex_matrix: numpy.ndarray[numpy.float32[m, 3]] = array([], shape=(0, 3), dtype=float32), face_matrix: numpy.ndarray[numpy.int32[m, 3]] = array([], shape=(0, 3), dtype=int32), v_normals_matrix: numpy.ndarray[numpy.float32[m, 3]] = array([], shape=(0, 3), dtype=float32), f_normals_matrix: numpy.ndarray[numpy.float32[m, 3]] = array([], shape=(0, 3), dtype=float32)) Invoked with: kwargs: vertex_matrix=TrackedArray([[ 0.0116172 , 0.35717618, 0.03688662], [ 0.01200351, 0.3561042 , 0.0396708 ], [ 0.01240174, 0.35544643, 0.04137919], ..., [-0.06300603, -0.34894037, 0.04522913], [-0.06259257, -0.34896928, 0.04515301], [-0.06218336, -0.34902242, 0.04501301]]), v_color_matrix=TrackedArray([[1., 1., 1., 1.], [1., 1., 1., 1.], [1., 1., 1., 1.], ..., [1., 1., 1., 1.], [1., 1., 1., 1.], [1., 1., 1., 1.]])

I guess you didn't finish 'test_02_model.py' well . It looks empty mesh.

The process of reconstruction of mesh was done with 2K2K. If there is a problem in the mesh session, you can check out this repository https://github.com/SangHunHan92/2K2K. We will update a version of the code when we are done with the changes. Thanks for the suggestion, we will pay attention to this in future updates.

[sunwonlikeyou]

You

I ran the all pre-processing codes. And finally running the {sh_file_name}, but 3d result looks weired. Especially 0-4 views textures are white(or none). and below result is generate_image = init_image.convert("RGBA") from generate_texture.py, line 295, view idx==0. I think initial texture has to have textures from input image. Is this weird? or right?

Thank you very much for supporting our work. You can check if the 0-4 views images are generated properly. We will be working on optimizing our pipeline in the near future. we will update a version of the code when we are done with the changes.

Thank you for reply. Does that mean 'inpainted' image? view 0 looks same above. But view 1-4 (below is view 1) are fine. But I'll check view0 again.

[admg67]

Thank you, this issue has been solved.

I ran the all pre-processing codes. And finally running the {sh_file_name}, but 3d result looks weired. Especially 0-4 views textures are white(or none). and below result is generate_image = init_image.convert("RGBA") from generate_texture.py, line 295, view idx==0. I think initial texture has to have textures from input image. Is this weird? or right?

Hello, the following problems occurred during the running of my code, can you help me solve them? (ultraman) root@I1c915cb88800301a87:/hy-tmp/Ultraman# ./bash/run_mesh.sh /hy-tmp/Ultraman/Ult 008 ./result : invalid optionsh: line 1: set: - set: usage: set [-abefhkmnptuvxBCHP] [-o option-name] [--] [arg ...] ./bash/run_mesh.sh: line 2:

′

\r
′

: c o m m a n d n o t f o u n d .

/

b a s h

/

r u

n
m

e s h . s h : l i n e 5 : '\r': command not found ./bash/run_mesh.sh: line 6: $'\r': command not found Start Loading Model ... Model Loaded !! 0%| | 0/1 [00:00<?, ?it/s]2.86102294921875e-06 /usr/local/miniconda3/envs/ultraman/lib/python3.8/site-packages/torch/nn/functional.py:4277: UserWarning: Default grid_sample and affine_grid behavior has changed to align_corners=False since 1.3.0. Please specify align_corners=True if the old behavior is desired. See the documentation of grid_sample for details. warnings.warn( /usr/local/miniconda3/envs/ultraman/lib/python3.8/site-packages/torch/nn/functional.py:4215: UserWarning: Default grid_sample and affine_grid behavior has changed to align_corners=False since 1.3.0. Please specify align_corners=True if the old behavior is desired. See the documentation of grid_sample for details. warnings.warn( 3.6102426052093506 100%|██████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 1/1 [00:15<00:00, 15.56s/it] PyMeshLab 0.1.7 based on MeshLab 2020.12d 0%| | 0/1 [00:00<?, ?it/s]008.ply 0%| | 0/1 [00:00<?, ?it/s] Traceback (most recent call last): File "models/2K2K/test_03_poisson.py", line 57, in main(args.save_path, args.save_name) File "models/2K2K/test_03_poisson.py", line 34, in main ms = pymeshlab.Mesh(vertex_matrix = xyz, v_color_matrix=rgba) TypeError: init(): incompatible constructor arguments. The following argument types are supported: 1. pymeshlab.pmeshlab.Mesh(vertex_matrix: numpy.ndarray[numpy.float32[m, 3]] = array([], shape=(0, 3), dtype=float32), face_matrix: numpy.ndarray[numpy.int32[m, 3]] = array([], shape=(0, 3), dtype=int32), v_normals_matrix: numpy.ndarray[numpy.float32[m, 3]] = array([], shape=(0, 3), dtype=float32), f_normals_matrix: numpy.ndarray[numpy.float32[m, 3]] = array([], shape=(0, 3), dtype=float32)) Invoked with: kwargs: vertex_matrix=TrackedArray([[ 0.0116172 , 0.35717618, 0.03688662], [ 0.01200351, 0.3561042 , 0.0396708 ], [ 0.01240174, 0.35544643, 0.04137919], ..., [-0.06300603, -0.34894037, 0.04522913], [-0.06259257, -0.34896928, 0.04515301], [-0.06218336, -0.34902242, 0.04501301]]), v_color_matrix=TrackedArray([[1., 1., 1., 1.], [1., 1., 1., 1.], [1., 1., 1., 1.], ..., [1., 1., 1., 1.], [1., 1., 1., 1.], [1., 1., 1., 1.]])

I guess you didn't finish 'test_02_model.py' well . It looks empty mesh.

[sunwonlikeyou]

0 view texture is generated from 2K2K mesh. But 2K2K code doesn't make texture. so your generate_uv or other generate textures can't make colored texture. That's why my result was above like.

[Thesaltedfish-single]

@admg67 May I ask if you could share your solution? I found that except for view0 which is blank, all the perspectives I generated in the future were blank on the front. In the end, the model I generated had a blank front and a decent back effect #8