tf-siammask

SiamMask implementation with Tensorflow 2.

Install

pip install tf-siammask

Example

import numpy as np
import PIL.Image
import siammask

sm = siammask.SiamMask()

# Weight files are automatically retrieved from GitHub Releases
sm.load_weights()

# Adjust this parameter for the better mask prediction
sm.box_offset_ratio = 1.5

img_prev = np.array(PIL.Image.open('data/cat1.jpg'))[..., ::-1]
box_prev = np.array([[227, 184], [381, 274]])
img_next = np.array(PIL.Image.open('data/cat2.jpg'))[..., ::-1]

# Predicted box and mask images is created if `debug=True`
box, mask = sm.predict(img_prev, box_prev, img_next, debug=True)

Test data

	Previous frame	Next frame
File name	./data/cat1_with_box.jpg	./data/cat2.jpg
Image

Predicted mask for ./data/cat2.jpg

TODO

• Bounding-box regression
• Mask refinement network
• Pre-trained model for Tensorflow 2.0
• Training code
• Object tracking code

Reference

@inproceedings{wang2019fast,
    title={Fast online object tracking and segmentation: A unifying approach},
    author={Wang, Qiang and Zhang, Li and Bertinetto, Luca and Hu, Weiming and Torr, Philip HS},
    booktitle={Proceedings of the IEEE conference on computer vision and pattern recognition},
    year={2019}
}