Computers successfully trained to identify animals in photos. Researchers trained a deep neural network to classify wildlife species using 3.37 million camera-trap images of 27 species of animals obtained from five states across the United States. The model then was tested on nearly 375,000 animal images at a rate of about 2,000 images per minute on a laptop computer, achieving 97.6 percent accuracy -- likely the highest accuracy to date in using machine learning for wildlife image classification. Source 2v.
97.6% is not bad but most will see that as not acceptable unless its 99.9% because I'm sure the process doesn't come cheap.
Machine Learning with Neural Networks is a huge breakthrough in computing in our age.
We are witnessing the new age of computer learning in which through self-teaching computer can slowly identify pictures/videos which they were unable in the past few years.
They can even identify also actions in videos; i.e: bat swing, swimming, dog paddling, etc.
Actually, the model may only be predicting a specific set of pictures of the same properties. If the model is fed with other pictures with different properties the result may be different. My opinion.