Achromatic Single Metalens Imaging via Deep Neural Network

Abstract

Meta-optics are attracting intensive interest as alternatives to traditional optical systems comprising multiple lenses and diffractive elements. Among applications, single metalens imaging is highly attractive due to the potential for achieving significant size reduction and simplified design. However, single metalenses exhibit severe chromatic performance degradation arising from material dispersion and the nature of singlet optics, making them unsuitable for full-color imaging requiring achromatic performance. In this work, we propose and validate a deep learning-based approach to enhance full-color imaging quality in single metalens systems. Our developed deep learning networks computationally reconstruct raw imaging captures by effectively refocusing the red, green, and blue primary channels, eliminating chromatic aberration and vignetting, and enhancing resolution. Importantly, these improvements are achieved without requiring any hardware modifications to the metalens itself. Through comprehensive evaluations on diverse synthetic and real-world data sets captured under various environmental conditions and focusing distances, our approach consistently demonstrates significant enhancements in image quality. By providing a practical and simplified implementation, our method overcomes the inherent limitations of meta-optics and enables the realization of achromatic metalenses without complex engineering. By addressing key challenges in full-color imaging for single metalenses, this research enables new practical applications in photography, videography, and micrography via the easy integration of metalenses with commercial cameras.