Skip to content
Source code for paper "Layered Embeddings for Amodal Instance Segmentation"
Python Jupyter Notebook C++ C
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
.ipynb_checkpoints
.vscode
build
common
mrcnn add readme Apr 15, 2019
pycocotools add readme Apr 15, 2019
samples
.gitignore
README.md Update README.md Sep 18, 2019
batchEval.py
datasets.py remove dependency on pycocotools._mask; will use MATLAB version of ev… Jul 17, 2019
deprecated.py
embedding_model.py
embedding_shapes.ipynb
experiment.py remove dependency on pycocotools._mask; will use MATLAB version of ev… Jul 17, 2019
extract_weights.py
load_weights.py initial upload Apr 15, 2019
metrics_hist.py
model.py
mrcnn_shapes.ipynb
myPyAmodalEvalDemo.py
params.py
post_processing.py initial upload Apr 15, 2019
shapes.py
utils.py

README.md

Implementation for paper "Layered Embeddings for Amodal Instance Segmentation"

Note: the email associated with the paper is no longer valid. Please contact me at yanfengliux@gmail.com

Paper link: https://link.springer.com/chapter/10.1007/978-3-030-27202-9_9

Fig 1. Image and training ground truth: front class mask, occlusion class mask, front instance mask, occlusion instance mask

Fig 2. Network architecture

Fig 3. Failure case where there is a three-stack

Fig 4. Incomplete instance mask for the object at the bottom in the failure case

Fig 5. Mask R-CNN fails when there many objects of the same class and the bounding boxes are too crowded

You can’t perform that action at this time.