IMPASTO: Integrating Model-Based Planning with Learned Dynamics Models for Robotic Oil Painting Reproduction

1. 1Learn dynamics models from data to predict brushstroke effects
2. 2Integrate model-based planning with learned dynamics models for multi-step stroke planning
3. 3Use force-sensitive control of deformable tools to execute stroke trajectories, forces, and colors
4. 4Train and test the IMPASTO system using a sequence of target oil painting images

Robotics engineers and AI researchers on a team can benefit from IMPASTO as it enables robotic systems to learn and reproduce complex oil paintings without human demonstrations or faithful simulators. This can be particularly useful in applications where human-robot collaboration is required

💡 IMPASTO enables robots to infer and execute stroke trajectories, forces, and colors needed to reproduce oil paintings without human demonstrations

🤖💡 IMPASTO: Robotic oil painting reproduction using model-based planning & learned dynamics models