ymgw55 / segment-anything-edge-detection

Unofficial edge detection implementation using the Automatic Mask Generation (AMG) of the Segment Anything Model (SAM).

segment-anything-edge-detection

This repository provides code for performing edge detection using the Automatic Mask Generation (AMG) of the Segment Anything Model (SAM). Since the code used in the paper is not currently available to the public, this implementation is based on the descriptions provided in the paper.

The image on the left is taken from the BSDS. The middle is the ground truth edge. The image on the right is the result of applying edge detection.

Requirements

This repository assumes that you can already use a SAM model. Note that this repository uses opencv-contrib-python, not opencv-python, so install it as follows:

pip install opencv-contrib-python

See the description for more details.

You will also need to download the model for Edge NMS beforehand.

cd /working/model
wget https://cdn.rawgit.com/opencv/opencv_extra/3.3.0/testdata/cv/ximgproc/model.yml.gz

Run

To generate the image above, do the following:

python example.py

The output result is generated in output/pred/example.

Assuming the BSDS path is /working/data/BSR_bsds500/BSR/, to generate output for the bsds500 test-set, do the following:

python test.py

The output result is generated in output/pred/${pp_name}/test.

Evaluation

We use py-bsds500 for edge detection. Some bugs have been fixed and ported to the py-bsds500 directory. Compile the extension module with:

cd py-bsds500
python setup.py build_ext --inplace

Then run:

# This will take about 3.5 hours.
python evaluate.py /working/data/BSR_bsds500/BSR/ ../output/pred/${pp_name} test --thresholds 99

Todo

• Faster calculation of the evaluation through parallel processing.

Reference

The code in this repository mainly uses code from the following two repositories. Thank you.

• segment-anything
• py-bsds500