CenterMask : Real-Time Anchor-Free Instance Segmentation

We propose a simple yet efficient anchor-free instance segmentation, called CenterMask, that adds a novel spatial attention-guided mask (SAG-Mask) branch to anchorfree one stage object detector (FCOS[33]) in the same vein with Mask R-CNN [9] Plugged into the FCOS object detector, the SAG-Mask branch predicts a segmentation mask on each detected box with the spatial attention map that helps to focus on informative pixels and suppress noise. We also present an improved backbone networks, VoVNetV2, with two effective strategies: ( I) residual connection for alleviating the optimization problem of larger VoVNet and (2) effective Squeeze-Excitation (eSE) dealing with the channel information loss problem of original SE. With SAG-Mask and VoVNetV[19], we deign CenterMask and CenterMask-Lite that are targeted each to large and small models, respectively. Using the same ResNet-101-FPN backbone, CenterMask achieves 38.3%, surpassing all previous state-of-the-art-methods while at a much faster speed. CenterMaskLite also outperforms the state-of-the-art by large margins at over 35fps on Titan Xp. We hope that CenterMask and VoVNetV2 can serve as a solid baseline of real-time instance segmentation and backbone network for various vision tasks, respectively. The Code is available at https:// github.com./youngwanLEE/CenterMask.