Camouflaged Instance Segmentation In-The-Wild: Dataset And Benchmark Suite

This paper pushes the envelope on camouflaged regions to decompose them into meaningful components, namely, camouflaged instances. To promote the new task of camouflaged instance segmentation in-the-wild, we introduce a new dataset, namely CAMO++, by extending our preliminary CAMO dataset (camouflaged object segmentation) in terms of quantity and diversity. The new dataset substantially increases the number of images with hierarchical pixel-wise ground-truths. We also provide a benchmark suite for the task of camouflaged instance segmentation. In particular, we conduct extensive evaluation of state-of-the-art instance segmentation methods on our newly constructed CAMO++ dataset in various scenarios. We also propose Camouflage Fusion Learning (CFL) framework for camouflaged instance segmentation to further improve the state-of-the-art performance. The dataset, model, evaluation suite, and benchmark will be publicly available at our project page. \url{https://sites.google.com/view/ltnghia/research/camo\_plus\_plus}