Lensless Camera

Lensless cameras replace traditional lenses with coded apertures and computational algorithms, aiming to create smaller, cheaper, and more versatile imaging systems. Current research heavily utilizes deep learning, particularly employing architectures like Vision Transformers (ViTs), generative diffusion models, and unrolled primal-dual networks, to reconstruct high-quality images from the raw, highly multiplexed sensor data. This approach addresses challenges like biofouling, inconsistent illumination, and low image resolution, enabling applications in diverse fields such as underwater monitoring, privacy-preserving face recognition, and even human pose estimation. The resulting compact and potentially low-cost devices offer significant advantages for various applications and are driving innovation in computational imaging.