A comparison of visual features used by humans and machines to classify wildlife

Abstract

In our quest to develop more intelligent machines, knowledge of the visual features used by machines to classify objects shall be helpful. The current state of the art in training machines to classify wildlife species from camera-trap data is to employ convolutional neural networks (CNN) encoded within deep learning algorithms. Here we report on results obtained in training a CNN to classify 20 African wildlife species with an overall accuracy of 87.5% from a dataset containing 111,467 images. We then used a gradient-weighted class-activation-mapping (Grad-CAM) procedure to extract the most salient pixels in the final convolution layer. We show that these pixels highlight features in particular images that are in most, but not all, cases similar to those used to train humans to identify these species. Further, we used mutual information methods to identify the neurons in the final convolution layer that consistently respond most strongly across a set of images of one particular species, and we then interpret the features in the image where the strongest responses occur. We also used hierarchical clustering of feature vectors (i.e., the state of the final fully-connected layer in the CNN) associated with each image to produce a visual similarity dendrogram of identified species. Finally, we evaluated how images that were not part of the training set fell within our dendrogram when these images were one of the 20 species \"known\" to our CNN in contrast to where they fell when these images were \"unknown\" to our CNN.