High-quality simultaneous bright-field and dark-field imaging based on the light-field matrix model

Dark-field imaging is widely used due to its high resolution and high contrast, but traditional methods are susceptible to noise interference and require frequent switching between bright-field and dark-field modes. This not only increases imaging complexity and sample exposure time but also affects imaging consistency. This paper proposes a simultaneous bright-field and dark-field imaging techniques based on the light-field matrix (LFM) model. Using a single light source under weak scattering conditions, the technique constructs a light-field matrix model and employs the optical field decomposition principle from transmission matrix (TM) theory to decompose the light field into three components: bright-field signals corresponding to single-scattered photons, dark-field signals corresponding to multiple-scattered photons, and noise signals. This method enables simultaneous acquisition of bright-field and dark-field images, simplifying the imaging process and reducing noise interference. Additionally, by incorporating low-rank and sparse matrix decomposition techniques, background light field interference is effectively eliminated, resulting in a 1.8-fold improvement in the dark-field imaging contrast. Experimental results demonstrate the broad applicability of this technique to various biological samples, validating its feasibility in the field of biological imaging and providing an innovative and practical solution for efficient and reliable imaging in complex scenarios.