Open full image in new tab
Zoom on image
Members remain the original copyright holder in all their materials here at Renderosity. Use of any of their material inconsistent with the terms and conditions set forth is prohibited and is considered an infringement of the copyrights of the respective holders unless specially stated otherwise.
Close to final testing with my OpenPose Poser plugin to export 3D poses to Stable Diffusion. On the upper left corner, we can see the 3D model posed in Poser, and then the plugin results next to it. We then drop this into ControlNet, disabled the OpenPose preprocessor, and then whatever character I prompt in SD will be in that pose, camera framing, and body proportions. In addition, if we look at the hand poses the AI has generated, they match the 3D character from Poser. This was my major motivation to create this plugin: to remove AI randomness from the poses and hands.
But don't take me wrong, this plugin helps a lot with achieving precise poses, but randomness can still mess things up depending on the prompt and the AI model used. Some AI models handle hands and poses better than others. It still takes some trial and error, but it's definitely better than the OpenPose's AI preprocessor - especially with poses involving multiple people. The results speak for themselves. Character pose and camera framing established in Poser, rendered in SD, postwork in Paintshop.