Detection of lameness and determination of the affected forelimb in horses by use of continuous wavelet transformation and neural network classification of kinematic data.

This study researches the use of continuous wavelet transformation and neural network classification to detect lameness in a horse’s forelimb. The results indicate that this method can successfully distinguish between a healthy horse and a lame one by analyzing the movement of the horse’s head and one forelimb.

Study Methodology

• The study investigated the application of continuous wavelet transformation and neural network classification in discerning gait data for identifying forelimb lameness in horses.
• Twelve adult horses with mild forelimb lameness were investigated in this study. The positioning of the head and right forelimb foot, along with various other joints, were measured via kinematic analysis both before and after blocking the palmar digital nerves.
• The research obtained eight sets of recordings from horses without lameness, eight from horses with right forelimb lameness, and eight from horses with left forelimb lameness.
• These positions were processed using continuous wavelet transformation. The researchers constructed feature vectors from the transformed signals.
• A neural network was trained using data from six horses and subsequently tested on the remaining two horses in each category – the process was continued until each horse had been utilized twice for both training and testing processes.

Study Findings

• According to the results, the wavelet-transformed vertical position of the head and right forelimb foot had a significantly higher Correct Classification Percentage (CCP) at 85% compared to untransformed data which stood at 21% only.
• Interestingly, adding data from the fetlock, carpal, or elbow joints did not contribute to an increase in the CCP above that for the head and foot alone.

Study Conclusion

• In their conclusion, the researchers find the wavelet transformation of gait data to be particularly insightful in detecting and differentiating forelimb lameness of horses.
• Moreover, all the required information to detect lameness and discern the side of the lameness can be gained by observing vertical head movement in tandem with the movement of the foot of a single forelimb.