Is it actaually the new standard to post missleading videos on websites and twitter where the phone captures the videos and also processes the reconstruction live, while this repo and the code make clear that this is not a viable way right now?!
The heavy reconstruction path expects external COLMAP/MASt3R/hloc camera solving plus server/workstation inference, then produces .ply/.pt/.glb outputs.
Just my 5 cents for the future!
Is it actaually the new standard to post missleading videos on websites and twitter where the phone captures the videos and also processes the reconstruction live, while this repo and the code make clear that this is not a viable way right now?!
The heavy reconstruction path expects external COLMAP/MASt3R/hloc camera solving plus server/workstation inference, then produces .ply/.pt/.glb outputs.
Just my 5 cents for the future!