This model is an instance segmentation network for 80 classes of objects. Mask R-CNN with Oct0.5ResNet50 backbone, FPN, light-weight RPN, SERes detection head, and dual attention segmentation head.
|MS COCO val2017 box AP||32.99%|
|MS COCO val2017 mask AP||28.37%|
|Max objects to detect||100|
Average Precision (AP) is defined and measured according to the standard MS COCO evaluation procedure.
im_data, shape: [1x3x480x480] - An input image in the format [1xCxHxW]. The expected channel order is BGR.
im_info, shape: [1x3] - Image information: processed image height, processed image width and processed image scale with respect to the original image resolution.
classes, shape: [100, ] - Contiguous integer class ID for every detected object,
0for background, that is, for no object
scores: shape: [100, ] - Detection confidence scores in the range [0, 1] for every object
boxes, shape: [100, 4] - Bounding boxes around every detected objects in the (top_left_x, top_left_y, bottom_right_x, bottom_right_y) format
raw_masks, shape: [100, 81, 14, 14] - Segmentation heatmaps for all classes for every output bounding box
[*] Other names and brands may be claimed as the property of others.