An Effective CNN and Transformer Fusion Network for Camouflaged Object Detection

Authors: Dongdong Zhang, Chunping Wang, Huiying Wang, Qiang Fu, Zhaorui Li.

1. Preface

• This repository provides code for "An Effective CNN and Transformer Fusion Network for Camouflaged Object Detection"

2. Proposed Baseline

2.1. Training/Testing

The training and testing experiments are conducted using PyTorch with a single NVIDIA GeForce RTX 3090 GPU of 24 GB Memory.

1. Configuring your environment (Prerequisites):

   • Creating a virtual environment in terminal: conda create -n CTFNet python=3.7.

   • Installing necessary packages: pip install -r requirements.txt.

2. Downloading necessary data:

   • downloading testing dataset and move it into ./data/TestDataset/, which can be found in this download link (Google Drive).

   • downloading training dataset and move it into ./data/TrainDataset/, which can be found in this download link (Google Drive).

   • downloading pretrained weights and move it into ./checkpoints/CTF-Net/Net_epoch_best.pth, which can be found in this [download link (Baidu Pan)](https://pan.baidu.com/s/1id9-vZy3bReN90PDXKfu4Q hyu2).

   • downloading Res2Net weights and move it into ./models/res2net50_v1b_26w_4s-3cf99910.pthdownload link (Google Drive).

   • downloading PVTv2 weights and move it into ./pvt_v2_b2.pth[download link (Baidu Pan)](https://pan.baidu.com/s/1n5d-q4Wj3EN7kLxNv6Xg1A a8of).

3. Training Configuration:

   • Assigning your costumed path, like --train_save and --train_path in etrain.py.

4. Testing Configuration:

   • After you download all the pre-trained model and testing dataset, just run etest.py to generate the final prediction map: replace your trained model directory (--pth_path).

2.2 Evaluating your trained model:

Assigning your costumed path, like pred_root and model_lst in MyEval.py.

Just run MyEval.py to evaluate the trained model.

pre-computed maps of CTF-Net can be found in [download link (Baidu Pan)](https://pan.baidu.com/s/1IPxvXD49Oq5yyjhNyiRqpw irmx).