Automatic Pain Detection on Horse and Donkey Faces

Abstract

Recognition of pain in equines (such as horses and donkeys) is essential for their welfare. However, this assessment depends solely on the ability of the observer to locate visible signs of pain since there is no verbal communication. The use of Grimace scales is proven to be efficient in detecting pain but is time-consuming and also dependent on the level of training of the annotators and, therefore, validity is not easily ensured. There is a need for automation of this process to help training. This work provides a system for pain prediction in horses, based on Grimace scales. The pipeline automatically finds landmarks on horse faces before classification. Our experiments show that using different classifiers for different poses of the horse is necessary, and fusion of different features improves results. We furthermore investigate the transfer of horse-based models for donkeys and illustrate the loss of accuracy in automatic landmark detection and subsequent pain prediction.

Towards Machine Recognition of Facial Expressions of Pain in Horses

“…With the promising results from the manually annotated EquiFACS datasets [51,52], we investigated methods for automated recognition of horse facial AUs in still images [110]. Previous work has explored automated detection of keypoint-based facial expression information, but in a simplified form compared to, for example, EquiFACS [111][112][113]. In studies by the authors of [111,112], the method learned to classify the appearance of certain facial areas directly in terms of pain, meaning the presence of a specific AU, was not determined.…”

“…Previous work has explored automated detection of keypoint-based facial expression information, but in a simplified form compared to, for example, EquiFACS [111][112][113]. In studies by the authors of [111,112], the method learned to classify the appearance of certain facial areas directly in terms of pain, meaning the presence of a specific AU, was not determined. In Lu et al [111], the system was trained to recognize and score the intensity of nine different facial areas on a scale between 0 and 2.…”

“…In Lu et al [111], the system was trained to recognize and score the intensity of nine different facial areas on a scale between 0 and 2. These three approaches were developed for sheep, horses and donkeys and all rely on handcrafted feature representation of image patches (Histograms of Oriented Gradients), although [112] includes some experiments on CNN features. This is a robust and classical method, but it limits the range and precision of expressions that can be represented.…”

Sparse Neural Network for horse face detection in a Smart Riding Club Biometric System

“…Hummel et al [15] and Li et al [27] proposed using the face pattern as it is rich in information about the life of horses such as pain, disease and feelings.…”

“…The objective of Hummel et al [15] was to recognize the pain in equines. They suggested employing the HOG features and SVM for pose estimation and the SIFT, LBP, HOG and VGG16 features as well as SVM for pain recognition.…”