Automated contour extraction for light-sheet microscopy images of zebrafish embryos based on object edge detection...

Area of Science:

• Developmental Biology
• Biotechnology
• Microscopy Imaging

Background:

• Embryo contour extraction is crucial for quantitative morphology analysis and understanding development.
• Light-sheet microscopy enables in toto time-lapse imaging of embryos like zebrafish.
• Extracting embryo contours from light-sheet microscopy data is challenging due to large data volumes and object variability.

Purpose of the Study:

• To develop a workflow for extracting zebrafish embryo contours without manual labeling.
• To utilize an edge detection method based on change point detection for contour extraction.
• To evaluate the performance and robustness of the proposed method against established techniques.

Main Methods:

• A novel workflow employing edge detection with a change point detection approach for embryo contour extraction.
• Comparison of the proposed method with Sobel, Laplacian of Gaussian, adaptive threshold, Multi Otsu, and k-means clustering.
• Assessment of edge detection accuracy and noise robustness.

Main Results:

• The proposed method demonstrated superior edge detection accuracy compared to Sobel, LoG, adaptive threshold, Multi Otsu, and k-means methods.
• The workflow exhibited enhanced noise robustness over Multi Otsu and k-means clustering-based methods.
• The method proved effective for automated, small-scale contour extraction of zebrafish embryos.

Conclusions:

• The developed workflow provides an effective, label-free method for zebrafish embryo contour extraction.
• This approach offers a viable alternative when deep learning or other non-deep learning methods are not applicable.
• The method facilitates quantitative analysis of embryo morphology, aiding developmental process understanding.