PALMER: a method capable of parallel localization of multiple emitters for high-density localization microscopy

Area of Science:

• Biophysics
• Optical Microscopy
• Computational Imaging

Background:

• High-density emitter localization enhances temporal resolution in localization microscopy.
• Slow image analysis speed currently limits the widespread adoption of high-density methods.
• Achieving both high spatial and temporal resolution is crucial for advanced microscopy applications.

Purpose of the Study:

• To develop a high-density localization method that overcomes the speed limitations of existing techniques.
• To improve the image analysis speed for super-resolution microscopy without compromising localization accuracy.
• To enable faster acquisition of super-resolution images with high spatial fidelity.

Main Methods:

• Implemented a novel high-density localization method, PALMER, utilizing Graphics Processing Unit (GPU) parallel computation.
• Integrated multiple-emitter fitting algorithms with model selection via Bayesian Information Criterion (BIC).
• Compared PALMER's performance against the established SSM_BIC method and conventional sparse localization techniques.

Main Results:

• PALMER achieves localization accuracy comparable to the SSM_BIC method.
• PALMER demonstrates a speed improvement of over two orders of magnitude compared to previous methods.
• High-density localization using PALMER offers up to a ~14-fold speed gain for super-resolution image generation at Nyquist resolution.

Conclusions:

• PALMER significantly accelerates high-density emitter localization in microscopy.
• The method enhances the practical utility of high-density localization for faster super-resolution imaging.
• PALMER provides a viable solution for overcoming speed bottlenecks in advanced microscopy image analysis.