Hyperspectral imaging (HSI) offers detailed tissue analysis compared to traditional imaging, which is crucial for surgeries needing precise differentiation.
Current handheld optical systems face limitations in focal depth, making them less effective for operating room use.
By integrating a focus-tunable liquid lens and using deep reinforcement learning for video autofocusing, our new approach significantly outperformed traditional methods and received positive feedback from neurosurgeons in usability trials.