Factors driving the discovery and utilization of a newly available area by African elephants

For large mammals, area expansion is a key conservation measure to prevent species’ decline and extinction. Yet, its success depends on whether animals discover and later use these areas. Here, using GPS data, we investigated how herds of elephants detected and used an area made available to them after the removal of a fence. We studied the elephants’ behaviour before and after the fence removal, accounting for seasonal changes in movement patterns. In contrast to previous studies, herds visited the newly available area within a month of the fence removal, and the maximum distance they moved into the new area was reached between 5 and 9 months after the fence removal. Yet, elephants did not preferentially visit the new area at night. By the second year, all herds had shifted their seasonal home ranges and incorporated the new area, in contrast to a previous range expansion event. Our analyses show that the regular proximity of elephants to the original fence, and the fact that the new area was generally composed of preferred habitats of the elephants, probably explained the rapid discovery and use of the area. Our study improves our understanding of animal exploration and the role of habitat quality, and thus may improve range expansion and corridor programmes.

Real-world fence removal from a single-image via deep neural network

At public space such as a zoo and sports facilities, the presence of fence often annoys tourists and professional photographers. There is a demand for a post-processing tool to produce a non-occluded view from an image or video. This “de-fencing” task is divided into two stages: one is to detect fence regions and the other is to fill the missing part. For a decade or more, various methods have been proposed for video-based de-fencing. However, only a few single-image-based methods are proposed. In this paper, we mainly focus on single-image fence removal. Conventional approaches suffer from inaccurate and non-robust fence detection and inpainting due to less content information. To solve these problems, we combine novel methods based on a deep convolutional neural network (CNN) and classical domain knowledge in image processing. In the training process, we are required to obtain both fence images and corresponding non-fence ground truth images. Therefore, we synthesize natural fence image from real images. Moreover, spacial filtering processing (e.g. a Laplacian filter and a Gaussian filter) improves the performance of the CNN for detecting and inpainting. Our proposed method can automatically detect a fence and generate a clean image without any user input. Experimental results demonstrate that our method is effective for a broad range of fence images.