Does synthetic animation of artificial training data improve machine learning performance in identifying herpetofauna?

Department: 

Contact: 

First Name: 

Hillary

Last Name: 

McCauley

Content: 

Project Description

The goal with all of this is to compare whether synthetic "animation" of specimen images increases the ability of a Machine Learning model to recognize/identify live specimens in photos taken in the field by a deployed camera system. 
 
 

Undergraduate Contribution

1) Creating a training dataset using preserved specimens, software manipulation of specimen images, and after effects. 2) Developing Machine Learning algorithm 3) Carrying out statistical analysis on the results
 
 

Requirements and/or Application Instructions

Experience with machine learning, natural history collections, and imaging software. Contact Chris Evelyn, evelyn@ccber.ucsb.edu

Tags: