The introduction of a surgical microscope in neurosurgery increased the spectrum and safety of interventions.
The surgeon operates with both hands on the patient while also controlling the visualization system to provide the proper view.
Despite the benefits of the microscope, low ergonomics may lead to long-term effects on the musculoskeletal system. The bulky binocular system may limit the setup's flexibility, hindering robotics integration into the operating theatre.

Recently, the introduction of the exoscopes overcame those limitations by enabling a flexible and ergonomic working environment, combined with improved image quality.
The exoscope is fixed on the surgical field through pneumatic arm holders that the surgeon can move manually or through foot control.
Although providing the surgeon with enhanced vision, the complexity of its use has emerged as a limiting factor.
Visual servoing techniques have been primarily studied in the context of camera automation in minimally invasive laparoscopy \cite{gruijthuijsen2021robotic}. Its potential has become even more significant in the context of neurosurgery, as the only solution for tracking instruments is to employ vision sensors. However, visual servoing techniques for exoscope automation are currently limited due to open challenges in neurosurgical practices.

This work proposes a framework for an autonomous vision-guided camera holder that tracks and follows a selected surgical instrument based on a visual servoing technique.
We envision this solution to automate the exoscope in neurosurgery, providing intuitive control of the system and reducing surgeon workload and operating time.