Octopus Learning and Behavior
Backyard Brains Logo

Neuroscience for Everyone!

+1 (855) GET-SPIKES (855-438-7745)


items ()

Octopus Learning and Behavior

Hi, I’m Ilya Chugunov, a rising sophomore at UC Berkeley, majoring in Electrical Engineering and Computer Science. My hobbies include: hiking, archery, and pondering deeply about why my code doesn’t compile. As a member of the Backyard Brains fellowship this summer I will be studying the fascinating, and possibly unseen, behaviors of the California Two Spot Octopus (Octopus Bimaculoides).

Classifying and understanding animal behavior has been a hot-topic endeavor of biologists since even before the great Charles Darwin, but historically has very often been assays with pen on paper note taking. With recent developments in both the processing power of your average computer and the mathematics behind this processing, machine-learning has sprung up as a powerful tool to help us see beyond what our brains can comprehend. We now have the novel ability to find patterns in everything from the behaviors of lab rats or communication of cuttlefish chromatophores without having to introduce the arduous task of a scientist to list through thousands of frames of video.

With my project I plan to do exactly this, to collect a plethora of video data of Bimac Octopodes and use computer-vision and machine-learning techniques to extract behaviors from this video without biasing it via human intervention.

Below is an example of some of my work so far; using background subtraction, erosion, and Gaussian filtering I am able to transform a recording of a moving animal (in this case a squid) into an easily trackable contour entity, and then create a graph of its movement pattern (on the bottom).

 Additionally, I have a working algorithm to detect and quantify breathing patterns in a resting octopus. It transforms the video stream of the octopus (top) into a graph of detected motion vs frame number (center), which when then smoothed with the Blackman-Harris window function, allows us easily to deduce the time between breaths, in this case, 6.67 seconds (bottom).

I hope to continue to develop tools for octopus video analysis in this coming month and to discover more and more the hidden patterns in this beautiful animal’s behavior. I am especially excited to look into quantifying their problem-solving skills through various intellect challenging exercises (like letting them explore a maze!).

 


No Comments

No comments yet.

RSS feed for comments on this post.

Sorry, the comment form is closed at this time.