Images are routinely collected as part of studies in ecology, and tools such as camera trapping, aerial survey and remote sensing produce massive quantities of images. Presently, the capacity for ecologists to collect these images vastly outstrips their capacity to process them at scale. This is partly because of the ease and convenience of image collection, but also because the techniques used to automate their processing often require a high degree of familiarity with methods from the computer sciences. As a result, ecologists are often burdened with processing large image collections manually, increasing the logistical requirement to conduct these studies, and often limiting scope and inference. Deep-learning techniques have dramatically improved the state of the art in numerous fields of societal importance, including natural language processing, medical image analysis and drug discovery. In ecology, these techniques have recently been applied to perform automatic recognition of plants and animals and are revolutionizing how wildlife practitioners collect and assemble large image data sets. However, uptake of these methods by the broader ecology community has been slow, due in part to lack of awareness of the capacities of these methods and low coordination between independent research efforts. This session will cover the tools and techniques available to process images commonly collected during ecological research. We will discuss which techniques have proved effective for a breadth of image types, and look at some recent studies that have benefitted from the automation that deep-learning methods provide. By assembling both wildlife practitioners with access to large image collections and ecologists with skills in this area, we hope to prompt greater exchange of best practice and spur collaboration on this topic.