Machine Learning
Due to the nature of the work at Natural State, we often require ways to streamline and automate the processing of large amounts of data. Machine learning is a powerful tool that can be used to achieve this. This section of the documentation will provide a brief overview of the machine learning techniques that we use, and how we use them.
These sections are largely broken down into object detection and classification, and will be updated as we develop new techniques. We will also detail the training process, alongside the inference process. Throughout, we will provide reproducing code snippets, and links to the relevant repositories.
Species Classification
Species classification is the process of identifying the species of an animal from an image. This is a common task in the field of computer vision, and is often used in the context of camera trapping data. At Natural State, we use species classification to identify the species of animals in camera trap images. This is a key step in the process of analysing camera trap data, and is often used to generate species-specific abundance estimates.