tags:
- fastai
Andrew Alpha
This is a proof-of-concept version of our Artificial Intelligence model to classify images of bark and ambrosia beetles. As an input, please use an image of a specimen, or a group of specimens, ideally in ethanol with a white background.
This proof-of-concept model has been trained on a preliminary sample of 12 species: Coccotypes dactyliperda, Hylesinus varius, Monarthrum fasciatum, Phloeosinus dentatus, Pityophthorus juglandis, Platypus cylindrus, Pycnarthrum hispidium, Scolytodes schwarzi, Xyleborinus saxesenii, Xyleborus affinis, Xylosandrus compactus, and Xylosandrus crassiusculus.
For correct interpretation of the results, it is important to consider not just the suggested name, but also the associated probability. Identification of other species is coming soon, as soon as they are added to the training set.
You can find example photos here (new tab) and test them using Hugging Face space to our model.
Our model is based on the MaxViT architecture taht ahs been fpre-trained on imagenet and fine tuned on our own dataset.