Case Study: Johns Hopkins University studies spider behavior faster with Appen AI data annotation

The Appen Data Annotation Platform enabled Johns Hopkins University researchers to precisely and quickly label data to train their models, track spider movements and understand underlying behavioral motivations

Johns Hopkins University’s behavioral neuroscience team studied how orb-weaving spiders build webs to better understand animal behavior and brain activity. The challenge was tracking every leg movement in dark, night-vision video and accurately annotating tens of thousands of frames so they could train machine vision models to analyze spider posture over time.

Using the Appen Data Annotation Platform, Johns Hopkins University quickly labeled 100,000 sampled frames and selected 10,000 high-quality annotations to train LEAP and DeepLabCut models. Appen’s annotations enabled the researchers to reduce a task that would have taken one person 1,500+ hours into a few weeks, and the resulting models achieved similar performance with mean pixel errors of 8.2 and 7.6, helping the team identify consistent web-building behavior patterns and advance future research.

Johns Hopkins University

Andrew Gordus

Assistant Professor of Biology