How can improve the model?

[rockmany2000]

I have a human photo, The fore result looks ok but the side result is incorrect. Is it an issue since the 4 multi-views are not enough?

[rockmany2000]

The photo result is fine in aiuni.ai.

[josepmy]

@rockmany2000 I detected the same issue, Did you find some solution or idea?

[rockmany2000]

I use SV3D to product 4 multi-views and feed to normal map model and create mesh. There is the mesh issue as Unique3D 4 multi-views.

[josepmy]

@rockmany2000 How you did it?

[rockmany2000]

@wukailu How about the project improvment? I found a new project which use 6 views in https://github.com/tencent/Hunyuan3D-1

[josepmy]

@rockmany2000 looks interesting, first testings are great. Thanks to share