[GUIDE] EVOVA 3D Neural Rendering Engine
EVOVA allows you to create 3D Gaussian Splatting data from images or videos.
To create a 3DGS item, capture the object, person, or place you want to reconstruct using a camera, smartphone, or other recording device. The captured subject should remain still during the entire capture process, and you should record it from many different angles.

For best results, we generally recommend using a video rather than individual photos. Video makes it easier to capture dense, continuous viewpoints with enough overlap between frames.
You can start from the left menu:
Create (Left Sidebar) → Create from Videos or Images (Open Create from Video & Image)
Good input data is the most important part of creating high-quality 3DGS content. Please follow the guidelines below when capturing your subject.
When capturing your subject, move slowly and cover the object from many angles.
For object capture, it is best to move around the object in a full 360-degree path. We recommend capturing from three different height levels:
This helps EVOVA understand the full shape and appearance of the object.
Avoid sudden camera movements, fast rotations, or shaking. These can cause motion blur and reduce reconstruction quality. Move slowly and steadily while recording.
If you are taking photos, make sure each image overlaps with the next one by at least 70%. If the gap between images is too large, the system may fail to match the images correctly, which can lead to reconstruction failure.
Lighting has a major impact on the final 3D result.
Avoid strong direct sunlight, as it can create harsh shadows. Ideal lighting conditions include:
Lighting should not change during capture. Avoid:
Consistent lighting helps produce cleaner and more realistic 3DGS results.