AI-Based Identification and Counting of Aedes Mosquito Eggs
Author Information
Author(s): Wang Minghao, Zhou Yibin, Yao Shenjun, Wu Jianping, Zhu Minhui, Dong Linjuan, Wang Dunjia
Primary Institution: East China Normal University, Shanghai, China
Hypothesis
Can AI improve the identification and counting of Aedes albopictus mosquito eggs in field surveys?
Conclusion
The AI model effectively identifies and counts Aedes albopictus mosquito eggs, showing great potential for enhancing vector control.
Supporting Evidence
- The model achieved a precision of 0.977 and recall of 0.978.
- The R-squared value between actual and identified egg counts was 0.997.
- The average detection time for a single tile was 0.48 seconds.
- The model was over 10 times faster than traditional counting methods.
- High accuracy was maintained even with complex backgrounds.
Takeaway
This study shows that a computer can quickly and accurately count mosquito eggs, which helps in controlling diseases like dengue.
Methodology
The study used a dataset of 1729 images and employed a two-stage Faster R-CNN model for detection and counting of mosquito eggs.
Potential Biases
Potential misidentification of impurities as mosquito eggs could affect accuracy.
Limitations
The model may struggle with high-density egg clusters and similar impurities that resemble mosquito eggs.
Participant Demographics
Wild strains of Aedes albopictus collected from the Minhang District of Shanghai, China.
Statistical Information
P-Value
0.977
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website