Video-rate processing in tomographic phase microscopy of biological cells using CUDA.

Area of Science:

• Biophysics
• Optical Microscopy
• Computational Imaging

Background:

• Tomographic phase microscopy (TPM) enables 3D refractive index (RI) mapping of biological samples.
• Current TPM reconstruction is computationally intensive, hindering real-time analysis and large dataset processing.
• Rapid 3D visualization and analysis are crucial for biological research and medical applications.

Purpose of the Study:

• To develop and implement a computationally efficient algorithm for rapid 3D RI map reconstruction from TPM data.
• To achieve reconstruction rates exceeding video rate for real-time 3D cell imaging.
• To enable advanced cellular parameter calculation and facilitate medical diagnostics.

Main Methods:

• Implemented novel phase extraction, phase unwrapping, and Fourier slice algorithms.
• Utilized parallel computing on a GPU (Nvidia Tesla K20c) with CUDA C.
• Acquired 73 interferometric projections from single-cell rotation over a 180° range.

Main Results:

• Achieved a 3D reconstruction rate exceeding 25 frames per second for 256x256 pixel interferograms.
• Successfully generated 64x64x64 voxel 3D RI maps.
• Demonstrated the capability to calculate additional cellular parameters concurrently with reconstruction.

Conclusions:

• The new implementation significantly accelerates 3D RI map reconstruction in TPM.
• This advancement enables real-time 3D cell visualization and processing.
• The technique holds promise for applications in medical diagnosis and cell sorting.