Abstract
Automated camera systems may focus on geometric tracking without considering aesthetic composition, while certain aesthetic evaluation software may lack the physical agency to move a camera to create an improved image. Disclosed are systems and methods for a closed-loop control architecture that can integrate a mobile robotic platform, such as a drone or wheeled robot, with a generative cognitive agent. Such a system can operate through an iterative cycle where an image can be captured, the agent can provide a natural language critique and a corrective command based on a high-level subjective goal, and a local controller can translate the command into physical actuation of the camera platform. This process can enable the system to iteratively refine a camera's physical position and orientation to satisfy an abstract, aesthetic goal, facilitating the autonomous discovery and capture of compositions based on high-level artistic direction.
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.
Recommended Citation
Sinha, Rohit; Save, Prathamesh; and Dwarakanath, Kumar N., "Autonomous Aesthetic Composition via a Generative Model in a Robotic Control Loop", Technical Disclosure Commons, (February 22, 2026)
https://www.tdcommons.org/dpubs_series/9368