Reversed Motion Capture...
Hello everyone, I'm newly registered here. I enjoy lurking and admiring your creations, but my skills aren't at a level to show off yet. But today I have a different topic that I need help with. I'm developing a project that will function much like motion capture software that uses markers. Thing is, I need to use it sort of reversed. I don't need all the functionality that commercial mocap solutions offer, but then on the other hand, need it to do a couple things that those solutions don't do. Basically, this is what I need in a nutshell;
Single camera detection of markers. Ideally, I would be able to take a single still frame capture and then have the software locate the markers within that still frame. From there I would like to establish a rigid body or numerous rigid bodies within the point cloud.
As I understand it, the commercial software usually has a "flex tolerance" that will alert the developer when a certain point has moved outside of the rigid body by a certain margin. My assumption is that the software is using some simple 3D vector tracking to accomplish this. This is part of the functionality that I need.
In fact, in the end, I would actually like to do things a little bit backward. I would like to digitally insert the markers into the still shots and have the software recognize them. As I understand it, the software typically knows the perspective of each individual camera and based on that information will detect point cloud movement and translate it into model animation. I want it to use the predefined vector geometry of the point cloud to tell me the perspective of the camera.
Any thoughts would be greatly appreciated. Thanks
Single camera detection of markers. Ideally, I would be able to take a single still frame capture and then have the software locate the markers within that still frame. From there I would like to establish a rigid body or numerous rigid bodies within the point cloud.
As I understand it, the commercial software usually has a "flex tolerance" that will alert the developer when a certain point has moved outside of the rigid body by a certain margin. My assumption is that the software is using some simple 3D vector tracking to accomplish this. This is part of the functionality that I need.
In fact, in the end, I would actually like to do things a little bit backward. I would like to digitally insert the markers into the still shots and have the software recognize them. As I understand it, the software typically knows the perspective of each individual camera and based on that information will detect point cloud movement and translate it into model animation. I want it to use the predefined vector geometry of the point cloud to tell me the perspective of the camera.
Any thoughts would be greatly appreciated. Thanks
Replies