Person-Re-ID-with-light-weight-network

An Improved Person Re-identification Method by light-weight convolutional neural network

---

we provide our training and testing code written in Keras for the paper "An Improved Person Re-identification Method by light-weight convolutional neural network".

1. This network is composed of Siamese architecture.
2. EfficientNet was employed to obtain discriminative features and reduce the demands for data.
3. we use x*y instead of (x-y)^2 as Square Layer.
4. We ran our tests on CUHK01 dataset.

Prerequisites

---

We used google colab to train and test the network:

• python version= 3.6.9

• keras version= 2.2.5

• GPU= Tesl P100

• GPU Memory= 16G

Evaluation

---

We arrived Rank@1= 70.1%, Rank@5= 95.2%, Rank@10= 99.1%, Rank@15= 99.1% and Rank@20= 99.2% with EfficientNetB0.

Citation

---

Please cite this paper in your publications if it helps your research:

@article{amouei2020transferlearning,
  title={An Improved Person Re-identification Method by light-weight convolutional neural network},
  author={Sajad Amouei Sheshkal and  Kazim Fouladi-Ghaleh and Hossein Aghababa},
  year={2020}
}