VGG 16 and Inception-v3 networks were trained using a texture dataset of muddied and clean cows. A new dataset with 600 images that is similar to the actual texture dataset was introduced and used to train the networks. The method used to train the networks was transfer learning. ImageNet weights were trained using the similar dataset, then the newly trained weights were trained again using the actual texture dataset which had 584 images. We used a novel CNN training method, which involved a middle training step training using transfer learning. The achieved validation accuracy was 95.5% which is considerably better than the state-of-the-art 87%.