Predicting Transition in Bean Beetle Embryo Development Using Wavelet Transforms and Neural Networks

Student Author(s)

Zachary Diener

Faculty Mentor(s)

Dr. Paul Pearson

Document Type

Poster

Event Date

4-15-2016

Abstract

As bean beetle embryos develop, time-lapse photographs of their eggs exhibit varying levels of brightness that correspond to different stages of maturation. These time signals can be analyzed to pinpoint when different stages occur. We have developed a method to accurately identify these changes in brightness using a combination of Haar Wavelet analysis and neural networks. We utilized the wavelet analysis to extract key features from the signal and then, using these features, we trained the neural network to pinpoint the transition points in the eggs’ development. We have studied these methods at various levels of noise using randomized situations. We are hoping our results support the usefulness of this method to analyze similar signals.

Comments

This project was supported by the Jacob E. Nyenhuis Student/Faculty Collaborative Summer Research Grant.

This document is currently not available here.

Share

COinS