mycosavant / fungi-recognition

Models for image-based fungi recognition. Best results in Kaggle FGVCx Fungi 2018. Supplementary material for WACV 2020 paper.

Fungi Recognition Models / Supplementary Material

This repository contains the models from our winning submission to the Kaggle FGVCx Fungi Classification Challenge and our WACV 2020 paper. If you use the models, please cite this paper.

---

  Fungi Recognition: A Practical Use Case.
  M. Sulc, L. Picek, J. Matas, T. Jeppesen, J. Heilmann-Clausen.
  2020 IEEE Winter Conference on Applications of Computer Vision (WACV), 2020.

---

In order to support research in fine-grained species classification and to allow full reproducibility of our results, we share the pre-trained Inception-v4 and Inception-ResNet-v2 CNN models trained for the FGVCx Fungi Classification Challenge, organized with the FGVC5 Workshop at CVPR 2018.

Usage

The models are shared in the form of TensorFlow checkpoints, and can be used directly within the TensorFlow slim framework. Information about the model parameters are described in the (above) WACV 2020 paper.

Models:

Inception Resnet V2 fine-tuned from ImageNet-pretrained checkpoint

Inception Resnet V2 fine-tuned from PlantCLEF-pretrained checkpoint

Inception V4 fine-tuned from ImageNet-pretrained checkpoint

Inception V4 fine-tuned from PlantCLEF-pretrained checkpoint

Inception V4 with double input size fine-tuned from ImageNet-pretrained checkpoint

Inception V4 with double input size fine-tuned from PlantCLEF-pretrained checkpoint

License

The models are licensed by the Creative Commons Attribution-NonCommercial 4.0 International License, i.e.: The models are shared only for non-commercial purposes. If you publish experiments/results based on the models, please attribute us by citing the (above) WACV paper.