Unifying Instance and Panoptic Segmentation with Dynamic Rank-1 Convolutions
This work provides a more efficient and accurate panoptic segmentation framework for computer vision researchers and practitioners by unifying instance and semantic segmentation within a single network.
This paper introduces DR1Mask, a fully-convolutional one-stage network that unifies instance and panoptic segmentation by adding a single classification layer for semantic segmentation. It achieves this through a novel dynamic rank-1 convolution (DR1Conv) module, resulting in a framework that is twice as fast as previous two-branch approaches for panoptic segmentation and 10% faster with 1 point higher mAP for instance segmentation compared to BlendMask.
Recently, fully-convolutional one-stage networks have shown superior performance comparing to two-stage frameworks for instance segmentation as typically they can generate higher-quality mask predictions with less computation. In addition, their simple design opens up new opportunities for joint multi-task learning. In this paper, we demonstrate that adding a single classification layer for semantic segmentation, fully-convolutional instance segmentation networks can achieve state-of-the-art panoptic segmentation quality. This is made possible by our novel dynamic rank-1 convolution (DR1Conv), a novel dynamic module that can efficiently merge high-level context information with low-level detailed features which is beneficial for both semantic and instance segmentation. Importantly, the proposed new method, termed DR1Mask, can perform panoptic segmentation by adding a single layer. To our knowledge, DR1Mask is the first panoptic segmentation framework that exploits a shared feature map for both instance and semantic segmentation by considering both efficacy and efficiency. Not only our framework is much more efficient -- twice as fast as previous best two-branch approaches, but also the unified framework opens up opportunities for using the same context module to improve the performance for both tasks. As a byproduct, when performing instance segmentation alone, DR1Mask is 10% faster and 1 point in mAP more accurate than previous state-of-the-art instance segmentation network BlendMask. Code is available at: https://git.io/AdelaiDet