The VietForest dataset

Our Vietnamese forestry dataset (VietForest) is a valuable resource comprising 156 classes of plant species predominantly found in the forests of Phu Tho province. This dataset encompasses a diverse range of flora, including 18 orders, 45 families, and 122 genera, totaling 26,251 images. The dataset showcases local plant species, many of which hold significant ecological importance and some of which are listed for conservation efforts. The dataset not only aids in understanding the local habitat but also provides detailed biological information associated with each species

The pictures in the dataset are taken in Phu Tho province in 2023.

Some pictures of our dataset are shown below:

Details

We trained multiple network structures, including ResNet50, DenseNet169, SENet-154, and Vision Transformer B-16, on two plant datasets using the loss functions described in the previous section. All networks were trained with the same configuration: a batch size of 32, 30 epochs, and an optimizer set to SGD with a learning rate of 0.001, momentum of 0.9, and weight decay of 5e-04. We utilized the StepLR learning rate scheduler with a step size of 10.

During the training process, we employed basic data augmentation techniques to enhance the robustness of our models. These techniques included random rotation and resizing of images to a standard size of 224x224 pixel

The detailed recognition results of VietForest dataset on PLANTCLEF2017 dataset and VietForest dataset are shown in the table below:

The detailed number of pictures per species:

Terms & Conditions of Use

The datasets are released for academic research only, and are free to researchers from educational or research institutes for non-commercial purposes.

Related Publications

All publications using VietForest or any of the derived datasets should cite the following papers:

Download

The requestor must sign in the commitment and send it to the database administrator (lan.lethi1@hust.edu.vn) by email.