Single-pixel microscopy is an imaging technique that forms images using a single detector instead of a camera. While powerful, versatile and allows simultaneous multidimensional image acquisition, it often requires many measurements, making imaging slow and inefficient. This work asks how measurement strategies can be optimized to reduce acquisition time without sacrificing image quality.

Augmented Reality (AR) glasses are rapidly evolving, but placing a camera directly in the user's field of view without blocking their vision remains a significant challenge. See-through cameras, such as the LightguideCam, solve this by hiding the camera mechanism inside a transparent prism. However, this compact design creates a new problem: internal light reflections cause severe blurring and complex visual artifacts in the captured images. The central question of this study was: How can we reconstruct crystal-clear images from this distorted sensor data in real-time, making the invisible camera truly usable?