Chameleo : walk like a chameleon detection with AI

Gofer, Shahar, Gris, Vanessa, Oren, Ariel, Sagi, Liran and van der Linden, Dirk (2021) Chameleo : walk like a chameleon detection with AI. In: ACI’21: Eigth International Conference on Animal-Computer Interaction. ACM, New York, US, pp. 1-5. ISBN 9781450385138

Preview

Text Gofer_et_al_2021_Chameleo_walk_like_a_chameleon_detection_with_AI.pdf - Accepted Version Download (4MB) | Preview

Abstract

Extensive research over the last few years has revealed the potential of computational based models to understand animal behavior. However, computational ethology using reptiles as models is still under investigation. Chameleons are known to have slow arboreal locomotion and present a distinct movement of rocking back-and-forth in between periods of the traditional quadrupedal walk. This curious, and yet, under-investigated behavior known as “leaf movement”, has been observed in different species of the genus Chamaeleo. Here we present our work-in-progress and propose the means to quantitatively examine plausible gaits of chameleons using an Artificial Neural Network system named Chameleo. We recorded and labeled around 8 hours of chameleons moving horizontally on a rope in an experimental setup and aim to use this data for training and further testing of the Neural Network. We expect that Chameleo will be an accurate and reliable model for the identification and classification of chameleon locomotion. Furthermore, our long-term goals are to 1) adapt Chameleo to a wider range of lizard behaviors, 2) make the model available for the scientific community through a website where researchers will be able to add additional models and datasets to further explore reptile behavior, 3) contribute to the welfare of pet chameleons, and finally 4) encourage citizen science and thus conservation and environmental protection of the species.

Actions (login required)

View Item

Downloads