Related Experiment Video
Updated: Jun 28, 2026
08:34
Automated Charting of the Visual Space of Housefly Compound Eyes
Published on: March 31, 2022
Automatic recognition and measurement of butterfly eyespot patterns
Margarida Silveira1, Antónia Monteiro
1Department of Electrical and Computer Engineering, IST University, Av. Rovisco Pais, 1049-001 Lisbon, Portugal. msilveira@isr.ist.utl.pt
Bio Systems
|October 29, 2008
Summary
Researchers developed an automated tool to detect and measure butterfly eyespot patterns using machine learning. This software accurately identifies and quantifies these crucial wing patterns in digital images.
Area of Science:
- Evolutionary biology
- Developmental biology
- Behavioral ecology
- Image analysis
- Machine learning
Background:
- Butterfly eyespots are key elements in evolutionary and developmental biology research.
- Automated tools are needed to keep pace with accelerating research on these patterns.
Purpose of the Study:
- To develop a tool for automatic detection and measurement of butterfly eyespot patterns in digital images.
- To create a machine learning algorithm for classifying eyespot and non-eyespot regions.
Main Methods:
- Utilized a machine learning algorithm trained on labeled images (eyespot vs. non-eyespot).
- Employed features based on circularity and symmetry for eyespot detection.
- Applied a 1D Hough Transform for measuring color ring radii after detection.
Main Results:
- Successfully trained software to recognize eyespot patterns in the nymphalid butterfly Bicyclus anynana.
- The developed algorithm demonstrated successful detection of eyespots in other butterfly species.
- The tool automatically detects eyespots and measures their ring radii.
Conclusions:
- The developed automated tool effectively detects and measures butterfly eyespots.
- This software can accelerate research in evolutionary and developmental biology by providing quantitative data.
- The algorithm shows potential for broader application across different butterfly species.
