Influence of the phase effect on gradient-based and statistics-based focus measures in bright field microscopy

Area of Science:

• Microscopy
• Biotechnology
• Image Analysis

Background:

• Autofocusing is critical for high-throughput microscopy and live-cell imaging.
• Reliable focus measures are essential for accurate image acquisition.
• Phase objects, like single cells, present challenges in bright-field microscopy due to low contrast at optical focus.

Purpose of the Study:

• To investigate the behavior of common focus measures with phase objects in microscopy.
• To analyze the impact of cell homogeneity on focus measure performance.
• To understand the influence of the phase effect on autofocusing algorithms.

Main Methods:

• Evaluation of gradient-based and statistics-based focus measures.
• Analysis of focus curves for homogeneous and inhomogeneous cell distributions.
• Interpretation using the phase effect equation and patch-wise analysis.
• Assessment of expert preferences in focus position selection.

Main Results:

• Focus measures exhibit untypical behavior (local minima instead of maxima) for homogeneous phase objects at optical focus.
• Gradient-based measures perform typically for inhomogeneous cells, while statistics-based measures show similar behavior to homogeneous cells.
• Bioprocess engineering experts preferred defocused images, indicating a subjective influence of the phase effect.

Conclusions:

• Standard autofocusing measures are unreliable for phase objects due to the phase effect.
• Cell distribution (homogeneity) significantly impacts focus measure performance.
• The phase effect influences both automated focus detection and human expert perception.