README.md

PlantCamo Dataset is the first dataset dedicated for plant camouflage detection. It contains over 1,000 images with plant camouflage characteristics.

Demo

https://github.com/yjybuaa/PlantCamo/assets/39208339/766645f1-6951-4cf0-baf1-8a5fb4b13bd9

Download the Dataset and Results

PlantCamo-full(Code: wnyc) Google drive link

PlantCamo-Train-and-Test(Code: hq87) Google drive link

Results(Code: 6o76) Google drive link

Usage

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

Download pvt_v2_b2.pth at here (Code: gy87) or Google drive link, and put it into .\pretrained_pvt

Train

Download PlantCamo-Train-and-Test at here(Code: hq87) or Google drive link, and put it into .\datasets

Test

Download trained model Net_epoch_best.pth at here(Code: b98f) or Google drive link, and put it into .\ckpt

Evaluation

You can find it in https://github.com/lartpang/PySODMetrics or you can run the metric_caller.py

Citation

We appreciate your support of our work!

@article{plantcamo,
      title={PlantCamo: Plant Camouflage Detection}, 
      author={Jinyu Yang and Qingwei Wang and Feng Zheng and Peng Chen and Aleš Leonardis and Deng-Ping Fan},
      journal={CAAI Artificial Intelligence Research (AIR)},
      year={2025}
}