Computer vision/object detection expert
Budget: $8 – $15 USD
Hello i am looking for help with a computer vision project. It is mostly available on a github page. But i need some modifications and additions to meet my needs. Would like to talk about it further thank you.
Application based on "Multiply" on gitHub.
Will need application to be able to work with sequences with 1 actor, as well as many actors. Currently I think it only works with 2+ actors.
Would prefer to just work with jpg sequences. Ffmpeg integration not needed.
Pre processing stage will need exporting camera and smpl-x to fbx. and importing camera/smpl-x from fbx is a must have. rather than forced trace input》training.
We can use example data from github for starters. I can also get data for single actor.
Vitpose plus open pose for preprocessing needed for added accuracy.
Fbx import and fbx export will be preferred method for camera and smpl-x.
Custom SMPL-X Maya rig and cameras will be handled on my side. I will provide data for this. If trace camera and smpl-x data can be exported to me as fbx (from pre processing stage). I can set up the system on my side to match so my data can be parsed into the trainer.
Application based on "Multiply" on gitHub.
Will need application to be able to work with sequences with 1 actor, as well as many actors. Currently I think it only works with 2+ actors.
Would prefer to just work with jpg sequences. Ffmpeg integration not needed.
Pre processing stage will need exporting camera and smpl-x to fbx. and importing camera/smpl-x from fbx is a must have. rather than forced trace input》training.
We can use example data from github for starters. I can also get data for single actor.
Vitpose plus open pose for preprocessing needed for added accuracy.
Fbx import and fbx export will be preferred method for camera and smpl-x.
Custom SMPL-X Maya rig and cameras will be handled on my side. I will provide data for this. If trace camera and smpl-x data can be exported to me as fbx (from pre processing stage). I can set up the system on my side to match so my data can be parsed into the trainer.