On-chip TIRF nanoscopy by applying Haar wavelet kernel analysis on intensity fluctuations induced by chip illumination

TL;DR

This paper introduces Haar wavelet kernel analysis (HAWK) to preprocess on-chip TIRF images, reducing artifacts from non-uniform illumination and enhancing super-resolution imaging with SOFI.

Contribution

The novel application of HAWK analysis to improve on-chip TIRF nanoscopy by mitigating illumination artifacts and enhancing super-resolution capabilities.

Findings

Resolution improved with HAWK and SOFI combination

Artifacts from non-uniform illumination reduced

Spatio-temporal sparsity increased in processed images

Abstract

Photonic-chip based TIRF illumination has been used to demonstrate several on-chip optical nanoscopy methods. The sample is illuminated by the evanescent field generated by the electromagnetic wave modes guided inside the optical waveguide. In addition to the photokinetics of the fluorophores, the waveguide modes can be further exploited for introducing controlled intensity fluctuations for exploitation by techniques such as super-resolution optical fluctuation imaging (SOFI). However, the problem of non-uniform illumination pattern generated by the modes contribute to artifacts in the reconstructed image. To alleviate this problem, we propose to perform Haar wavelet kernel (HAWK) analysis on the original image stack prior to the application of (SOFI). HAWK produces a computational image stack with higher spatio-temporal sparsity than the original stack. In the case of multimoded non-uniform illumination patterns, HAWK processing bre aks the mode pattern while introducing spatio-temporal sparsity, thereby differentially affecting the non-uniformity of the illumination. Consequently, this assists nanoscopy methods such as SOFI to better support super-resolution, which is otherwise compromised due to spatial correlation of the mode patterns in the raw image. Furthermore, applying HAWK prior to SOFI alleviates the problem of artifacts due to non-uniform illumination without degrading temporal resolution. Our experimental results demonstrate resolution enhancement as well as reduction in artifacts through the combination of HAWK and SOFI.