Use Case and High-Level Description

This model is an instance segmentation network for 80 classes of objects. It is a Mask R-CNN with EfficientNet-B2 backbone, light-weight FPN, RPN, detection and segmentation heads.



Metric Value
MS COCO val2017 box AP 35.0%
MS COCO val2017 mask AP 31.2%
Max objects to detect 100
GFlops 29.334
MParams 13.5673
Source framework PyTorch*

Average Precision (AP) is defined and measured according to standard MS COCO evaluation procedure.


  1. name: image , shape: [1x3x608x608] - An input image in the format [1xCxHxW]. The expected channel order is BGR.


  1. name: labels, shape: [100] - Contiguous integer class ID for every detected object.
  2. name: boxes, shape: [100, 5] - Bounding boxes around every detected objects in (top_left_x, top_left_y, bottom_right_x, bottom_right_y) format and its confidence score in range [0, 1].
  3. name: masks, shape: [100, 28, 28] - Segmentation heatmaps for every output bounding box.

Legal Information

[*] Other names and brands may be claimed as the property of others.