2D to 3D Conversion Plugin for Blender

I’m a professional Stereoscopic Conversion Artist that has been developing a conversion plugin for Blender, and was looking to gauge the communities interest in a plugin of this nature, and network with some possible beta testers. This is an early test for the current demo I’m working on, to give you an idea of what I’m doing. All feedback is welcome, and greatly appreciated. Thanx.

Is your tool set a development of roto tools?

Hello tuffsuper, sounds interesting, hard to tell what the demo is doing, do you have any more details on what your plugin is aiming to achieve?

check in python forum to see if it does not already exist !

good luck

happy bl

I did a few stereo conversions in blender. Mostly by masking and z-separation of elements. Is the plugin using optical flow in any way (depth buildup?)? Because I think Blender lacks such tools to be proper stereo conversion tool. Something like Ocula, or Dimension.

We use the current masking and tracking tools, but aren’t necessarily developing the tools, just utilizing them to extract data for the meshes created using the plugin. The basic concept is that we import image or image sequences to planes, then create simple meshes on those planes and attach all the individual vertices to bones that are then constrained to the camera, each vertices can then be extruded in a 3D space to create depth. Then by using a stereoscopic camera rig you can render out left and right eyes. The masking and tracking tools are used primarily for image sequences.

The aim of the program is to convert a 2D source image into stereoscopic 3D, while not distorting or modifying the original 2D source image. And this is achieved by constraining each pixel to the camera, to maintain continuity no matter where the pixel is in “Z space” (X and Y being the pixels coordinates on the image.

I’m not familiar with Ocula or Dimension, but in my experience with conversion, simplicity is almost always the way to go. The plugin I’m developing is currently only for placing objects in a 3D space by projecting images onto user created meshes. And this doesn’t require a ton of new technology, we just need basic 3D tool sets, making Blender the perfect platform for conversion. Not that Blender doesn’t provide a robust tool suite, it just isn’t needed.

Here are screen grabs from my demo blend file, and one of the source images. Hopefully, this better demonstrates how the plugin works.




Some more screen shots, to demonstrate how the plugin works.




I can be a beta tester, ask me for anything you want

This sounds very interesting I look forward to seeing it