Unsupervised SAM segmentation of zebrafish body: Application to melanin analysis

Area of Science:

• * Developmental Biology
• * Toxicology
• * Biomedical Imaging

Background:

• * Zebrafish are valuable models for human health studies due to transparency and genetic similarity.
• * Accurate melanin quantification is crucial for assessing pigmentation disorder treatments and inhibitor efficacy.
• * Existing methods for melanin quantification in zebrafish can be labor-intensive and lack precision.

Purpose of the Study:

• * To develop a precise and efficient method for quantifying melanin in zebrafish.
• * To investigate the effects of 1-phenyl 2-thiourea (PTU) and α-Arbutin on melanin synthesis.
• * To leverage the Segment Anything Model (SAM) for automated zebrafish image segmentation.

Main Methods:

• * Employed the Segment Anything Model (SAM) for unsupervised image segmentation of zebrafish.
• * Achieved high accuracy (100%) in generating zebrafish body masks with SAM.
• * Calculated melanin content via pixel-level integration within segmented body masks.

Main Results:

• * Melanin synthesis was observed to be inhibited with increasing concentrations of PTU and α-Arbutin.
• * The SAM-based approach demonstrated high accuracy in mask selection (96.7%).
• * The proposed method provided quantitative results comparable to existing techniques.

Conclusions:

• * The Segment Anything Model (SAM) offers a precise and efficient tool for melanin quantification in zebrafish.
• * This automated approach simplifies the process and reduces manual effort in image analysis.
• * The findings confirm the inhibitory effects of PTU and α-Arbutin on zebrafish melanin synthesis.