S_T_Mamba: A Novel Jinnan Calf Diarrhea Behavior Recognition Model Based on Sequence Tree Mamba

Area of Science:

• Veterinary Medicine
• Animal Behavior
• Machine Learning

Background:

• Accurate recognition of calf diarrhea is vital for livestock health.
• Conventional methods struggle with subtle behavioral differences.

Purpose of the Study:

• To introduce a novel model for recognizing diarrhea-related behaviors in Jinnan calves.
• To enhance the precision of calf health monitoring through advanced AI.

Main Methods:

• Developed the S_T_Mamba (Sequence Tree Mamba) model.
• Utilized a sequence processing strategy for temporal dependencies.
• Incorporated a tree state space module (TreeSSM) for long-range pixel feature aggregation.

Main Results:

• S_T_Mamba achieved 99.78% accuracy in Jinnan calf diarrhea behavior recognition.
• The model outperformed existing methods by 0.59% to 1.99%.
• Demonstrated superior ability to distinguish between similar behavioral patterns.

Conclusions:

• The S_T_Mamba model offers a significant advancement in calf diarrhea behavior recognition.
• This technology can improve early detection and management of calf health issues.
• The model's effectiveness highlights the potential of advanced AI in precision livestock farming.