Here at the Department of Amphibological Research, we take image recognition on new expeditions in the natural world and tease out the limits of intelligence. We invite you to collaborate with our team by providing new inspiration for our amphibological experiments!


1. Discover a specimen

Do you think you may have found something new and unseen by the cloud?
Capture and compose it for further analysis by our team. Check out our previous experiments to see which pictures may lead to the most amphibological results.


2. Artificial analysis

Use some kind of image recognition AI to ‘interpret’ the image. Our current favorites are Demitasse (app) and Google Cloud Vision API (online). Other popular platforms such as Amazon Rekognition are popular online platforms (you should be able to ‘try’ their products for free, after registration). CamFind, Image Recognizer and TensorFlow Object Detection are quick and easy alternatives for smartphones… but feel free to get creative: test your own image recognition softwares, describe the picture using only emojis or translate it into sounds. The only rule is that no direct human-to-human communication is allowed!


3. Equivocal Interpretation

Once you have a description you are happy with, fill in the submission form and send it our team. Our team of amphibological artists will render the new specimen based on your description. Interesting findings will be added to the amphibological archive for collective inspiration.


The May 1922 cover of Scientific American, “with waterproof paints and canvas: Painting a submarine scene at first hand”. NOAA PHOTO LIBRARY/ CC BY 2.0