I need to train a neural network to recognize photos of leaves.

I have only laboratory photos with a white background, but I need leaf photos from nature.

Do you know where I can get this type of photos?

I need > 100 photos of each leaf (tree species).

  • Here is more info about this my project: BLOG POST
    – slinto
    Commented May 30, 2017 at 16:42

2 Answers 2


Leafsnap consists of images of leaves taken from two different sources, as well as their automatically-generated segmentations:
23147 Lab images, consisting of high-quality images taken of pressed leaves, from the Smithsonian collection. These images appear in controlled backlit and front-lit versions, with several samples per species.
7719 Field images, consisting of "typical" images taken by mobile devices (iPhones mostly) in outdoor environments. These images contain varying amounts of blur, noise, illumination patterns, shadows, etc.
The dataset currently covers all 185 tree species from the Northeastern United States.

Cleared Leaf Image Database is an online collection of cleared leaf images from different investigators and labs around the world gathered together in a single place.


There are some operational projects about plant recognition:

  • The Pl@ntNet project has a large database of plant images, partly built by volunteer contributions. They also have a companion mobile app for plant recognition. Have a look at their terms here to see how you could use these images: http://identify.plantnet-project.org/api/about/terms

  • PlantVillage Images is an open access database of plant, focused on crop diseases. Probably not a lot of trees...


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.