I am working on applying anipose to an maDLC project. I have an established pipeline for single animal projects already that uses anipose to filter and triangulate a 5 camera setup with labels in each frame to define the origin, x and y axes of our desired coordinate frame. I would like to apply the same/similar pipeline to an maDLC project with 4 cameras. please I have not been able to find any information on doing so in DLC or anipose documentation, and looking thru triangulate.py and other anipose code it seems it may require some light revision of the code base. My basic plan is:
1) Run analyze_videos and refine_tracklets in DLC, iterate until 2D tracking is satisfactory.
2) Place analyzed_video outputs into the directory structure expected by anipose (and maybe rename files to the anipose style if necessary?)
3) Run anipose from there. I anticipate that I would need to edit anipose code to differentiate between individuals for the multianimalbodyparts in all or some of filter_pose, filter3d, label*, and triangulate.py.
Is there already a solution to this? If not, does my plan sound right and could you provide any guidance on doing so?
I am working on applying anipose to an maDLC project. I have an established pipeline for single animal projects already that uses anipose to filter and triangulate a 5 camera setup with labels in each frame to define the origin, x and y axes of our desired coordinate frame. I would like to apply the same/similar pipeline to an maDLC project with 4 cameras. please I have not been able to find any information on doing so in DLC or anipose documentation, and looking thru triangulate.py and other anipose code it seems it may require some light revision of the code base. My basic plan is:
1) Run analyze_videos and refine_tracklets in DLC, iterate until 2D tracking is satisfactory. 2) Place analyzed_video outputs into the directory structure expected by anipose (and maybe rename files to the anipose style if necessary?) 3) Run anipose from there. I anticipate that I would need to edit anipose code to differentiate between individuals for the multianimalbodyparts in all or some of filter_pose, filter3d, label*, and triangulate.py.
Is there already a solution to this? If not, does my plan sound right and could you provide any guidance on doing so?
Thank you for your help in advance.