Is this project an undergraduate, graduate, or faculty project?

Undergraduate

group

What campus are you from?

Daytona Beach

Authors' Class Standing

Brian Danaher - Senior Matthew Brown - Senior Ying Zheng - Senior

Lead Presenter's Name

Brian Danaher

Faculty Mentor Name

Dr. Mihhail Berezovski

Abstract

An experiment was performed to investigate an alternative pooling method - Variable Stride (VS) - for use in convolutional neural networks. Three VS methods were compared to Maxpool and Avgpool in three different network configurations tasked with classifying diabetic retinopathy images between healthy retinas and retinas with advanced retinopathy. Each combination of network structure and pooling method was run multiple times, and the AUCs, losses, accuracies, as well as the training speed of each run were all collected. Two-tailed t-tests were then run on the prior metrics to quantify the relative performance of each pooling method in each scenario. It was found that VS consistently performs worse than Maxpool and Avgpool in this specific scenario across all metrics.

Did this research project receive funding support from the Office of Undergraduate Research.

No

Share

COinS
 

Evaluating the Variable Stride Algorithm for Identifying Diabetic Retinopathy

An experiment was performed to investigate an alternative pooling method - Variable Stride (VS) - for use in convolutional neural networks. Three VS methods were compared to Maxpool and Avgpool in three different network configurations tasked with classifying diabetic retinopathy images between healthy retinas and retinas with advanced retinopathy. Each combination of network structure and pooling method was run multiple times, and the AUCs, losses, accuracies, as well as the training speed of each run were all collected. Two-tailed t-tests were then run on the prior metrics to quantify the relative performance of each pooling method in each scenario. It was found that VS consistently performs worse than Maxpool and Avgpool in this specific scenario across all metrics.

 

To view the content in your browser, please download Adobe Reader or, alternately,
you may Download the file to your hard drive.

NOTE: The latest versions of Adobe Reader do not support viewing PDF files within Firefox on Mac OS and if you are using a modern (Intel) Mac, there is no official plugin for viewing PDF files within the browser window.