Multi-Stage Transfer Learning System with Lightweight Architectures in Medical Image Classification

Presentation Type

Poster

Student

Yes

Track

Health Care Application

Abstract

Transfer Learning is currently popular in Medical Image classification. Transfer Learning methods are extensively applied with CNN’s such as Res-net, Densenet, VGG16, Inception, etc. for various medical diagnoses. However, these models are computationally expensive and over parameterized. Another challenge we identified is limited labeled datasets are available in the medical image domain preventing the major advancements in Transfer Learning for Medical image classification. We propose a Multi-Stage Transfer Learning System using Lightweight Architecture to tackle limited target dataset problem with quicker training time. Preliminary results suggest that our model performed well on CT Head images by improving the accuracy over traditional single-stage transfer learning.

Start Date

2-11-2020 1:00 PM

This document is currently not available here.

Share

COinS
 
Feb 11th, 1:00 PM

Multi-Stage Transfer Learning System with Lightweight Architectures in Medical Image Classification

Volstorff A

Transfer Learning is currently popular in Medical Image classification. Transfer Learning methods are extensively applied with CNN’s such as Res-net, Densenet, VGG16, Inception, etc. for various medical diagnoses. However, these models are computationally expensive and over parameterized. Another challenge we identified is limited labeled datasets are available in the medical image domain preventing the major advancements in Transfer Learning for Medical image classification. We propose a Multi-Stage Transfer Learning System using Lightweight Architecture to tackle limited target dataset problem with quicker training time. Preliminary results suggest that our model performed well on CT Head images by improving the accuracy over traditional single-stage transfer learning.