YOLACT: Real-Time Instance Segmentation

@article{Bolya2019YOLACTRI,
  title={YOLACT: Real-Time Instance Segmentation},
  author={Daniel Bolya and Chong Zhou and Fanyi Xiao and Y. Lee},
  journal={2019 IEEE/CVF International Conference on Computer Vision (ICCV)},
  year={2019},
  pages={9156-9165}
}
  • Daniel Bolya, Chong Zhou, +1 author Y. Lee
  • Published 2019
  • Computer Science
  • 2019 IEEE/CVF International Conference on Computer Vision (ICCV)
  • We present a simple, fully-convolutional model for real-time instance segmentation that achieves 29.8 mAP on MS COCO at 33 fps evaluated on a single Titan Xp, which is significantly faster than any previous competitive approach. [...] Key Method We accomplish this by breaking instance segmentation into two parallel subtasks: (1) generating a set of prototype masks and (2) predicting per-instance mask coefficients. Then we produce instance masks by linearly combining the prototypes with the mask coefficients. We…Expand Abstract
    YOLACT++: Better Real-time Instance Segmentation
    • 10
    • PDF
    EOLO: Embedded Object Segmentation only Look Once
    Mask Encoding for Single Shot Instance Segmentation
    • 3
    • PDF
    CenterMask: Single Shot Instance Segmentation With Point Representation
    Efficient Instance Segmentation Network
    BlendMask: Top-Down Meets Bottom-Up for Instance Segmentation
    • 17
    • Highly Influenced
    • PDF
    Where are the Masks: Instance Segmentation with Image-level Supervision
    • 14
    • PDF
    PolarMask: Single Shot Instance Segmentation With Polar Representation
    • 41
    • Highly Influenced
    • PDF

    References

    Publications referenced by this paper.
    SHOWING 1-10 OF 57 REFERENCES
    Instance-Sensitive Fully Convolutional Networks
    • 239
    • PDF
    Instance-Aware Semantic Segmentation via Multi-task Network Cascades
    • 731
    • PDF
    Semantic Instance Segmentation with a Discriminative Loss Function
    • 178
    • PDF
    Box2Pix: Single-Shot Instance Segmentation by Assigning Pixels to Object Boxes
    • 23
    • Highly Influential
    • PDF
    InstanceCut: From Edges to Instances with MultiCut
    • 125
    • PDF
    R-FCN: Object Detection via Region-based Fully Convolutional Networks
    • 2,554
    • PDF
    SSD: Single Shot MultiBox Detector
    • 8,288
    • Highly Influential
    • PDF
    Straight to Shapes: Real-Time Detection of Encoded Shapes
    • 21
    • Highly Influential
    • PDF
    Segmentation-Aware Convolutional Networks Using Local Attention Masks
    • 68
    • PDF