Camera traps have become an extensively utilized tool in ecological research, but the manual processing of images created by a network of camera traps rapidly becomes an overwhelming task, even for small camera trap studies. We used transfer learning to create convolutional neural network (CNN) models for identification and classification. By utilizing a small dataset with an average of 275 labeled images per species class, the model was able to distinguish between species and remove false triggers. We trained the model to detect 17 object classes with individual species identification, reaching an accuracy up to 92% and an average F1 score of 85%. Previous studies have suggested the need for thousands of images of each object class to reach results comparable to those achieved by human observers; however, we show that such accuracy can be achieved with fewer images. With transfer learning and an ongoing camera trap study, a deep learning model can be successfully created by a small camera trap study. A generalizable model produced from an unbalanced class set can be utilized to extract trap events that can later be confirmed by human processors.

Point 1: Camera traps have become an extensively utilized tool in ecological research, but the processing of images created by a network of camera traps rapidly becomes an overwhelming task, even for small networks. Point 2: We used transfer training to create convolutional neural network (CNN) models for identification and classification. By utilizing a small dataset with less than 10,000 labeled images the model was able to distinguish between species and remove false triggers. Point 3: We trained the model to detect 17 object classes with individual species identification, reaching an accuracy of 92%. Previous studies have suggested the need for thousands of images of each object class to reach results comparable to those achieved by human observers; however, we show that such accuracy can be achieved with fewer images. Point 4: Additionally, we suggest several alternative metrics common to computer science studies to accurately evaluate the performance of such camera trap image processing models, as well as methods to adapt the model building process to two targeted purposes.