Deep-Learning Based Quantification of Bovine Oocyte Quality From Microscopy Images

Area of Science:

• Veterinary Science
• Biotechnology
• Artificial Intelligence in Animal Reproduction

Background:

• Bovine in vitro embryo production success is limited by oocyte quality.
• Current oocyte selection relies on subjective, time-consuming visual assessment by embryologists.
• Inconsistent oocyte evaluation impacts embryo development and production efficiency.

Purpose of the Study:

• To develop a semi-automatic, objective method for scoring immature bovine oocyte quality.
• To utilize deep learning for classifying oocyte competence.
• To improve the efficiency and success rate of bovine in vitro embryo production.

Main Methods:

• A deep learning model was developed for oocyte quality classification.
• The model was trained and validated using real-world images of immature bovine oocytes.
• Oocyte competence was labeled based on subsequent development into blastocysts after fertilization.

Main Results:

• The proposed deep learning model provides a robust and objective assessment of oocyte quality.
• The automated classification is more consistent than subjective visual evaluations by embryologists.
• The method's performance is comparable to existing approaches for human oocytes.

Conclusions:

• A semi-automatic, real-time method for scoring bovine immature oocytes using stereo-microscopy images has been established.
• This AI-driven approach significantly reduces the time and subjectivity in in vitro embryo production.
• The method promises to enhance the overall success rates of bovine embryo development.