Paired RGB images, camera parameters, and complete camera-frustum point clouds. The point clouds contain all available source-mesh vertices inside each image's camera frustum, including vertices hidden behind visible surfaces. They are variable-length XYZ arrays in the OpenCV camera frame (x right, y down, z forward). No fixed point count, sampling, normalization, or patch assignment is applied to the stored targets. - assetsmanifest.json: one record per image, with relative paths, point count, coordinate frame, image derivation modes, and SHA256 hashes. - shards/imgfullgtXXXX.tar: independent archives containing complete sample files, grouped by scene. Each archive is at most 5,000,000,000…
Independent publisher
Yao Mingrui
add2333
Models in Library0
Datasets in Library1
Models on Hugging Face—
Followers—
Compare the published evidence
This profile connects the 1 model and dataset record shown below to their upstream cards and sources. The library counts describe SAVRN's coverage, rather than this publisher's entire output.
Published tasks: Image to 3d.
Compare model specifications, source dates, and hardware estimates on the individual records. Publisher identity verification is separate from a model's quality, deployment readiness, or suitability for your work.