Towards high-performance polarimeters with large-area uniform chiral shells: a comparative study on the polarization detection precision enabled by the Mueller matrix and deep learning algorithm.

Polarization detection and imaging technologies have attracted significant attention for their extensive applications in remote sensing, biological diagnosis, and beyond. However, previously reported polarimeters heavily relied on polarization-sensitive materials and pre- established mapping relationships between the Stokes parameters and detected light intensities. This dependence, along with fabrication and detection errors, severely constrain the working waveband and detection precision.