Identification the face image of the sick sheep using visual geometry group (VGG) 19.

Sheep are livestock that ranks third after cows and goats that meet the needs of meat for the people of Indonesia, therefore it is very important to know the health condition of the sheep that will be consumed later. Sick sheep can be identified based on the sheep's facial expressions. In this study the authors applied VGG19 to identify sick sheep face. System performance is measured based on the value of accuracy, precision, recall, and F-Measure. After testing starting from 50 epochs to 200 with a learning rate (0.0001), and trying various optimizers such as Adamax, SGD (Stochastic Gradient Descent), and Adam, the results show an accuracy value of 89.4% and an average precision of 65.6%, recall 81% and f-measure 69.7%, with an mcc value of 0.44 using epoch 100, learning rate 0.0001, and the Adamax optimizer. These values are influenced by the dataset of the training image of 800 images, the validation image of 229 images, and the test image of 144 images. In testing for search the best side of the sheep's face image to be identified by the system, get the results that the sheep's face image from the front side gets an accuracy value of 94% with an average precision of 49%, recall 48% and f-measure 48% with an mcc value. -0.03. [ABSTRACT FROM AUTHOR]