Automated Ultrasound Technology to Diagnose Traumatic Retinal Detachment

Abstract

Our objective is to develop and validate a prototype mobile AI-enabled software platform that can accurately detect sonographic signs of retinal detachment and support a regulatory filing with the Food and Drug Administration. Our goal is to make the deep-learning ultrasound algorithms commercially available for use in the clinical environment to rapidly detect retinal detachment along with the status of macula to direct appropriate therapy. We hypothesized that machine learning algorithms will be reproducible and will have accurate diagnostic capability to detect findings specific to retinal detachment and macular involvement on ultrasound images. To date, we have successfully extracted, labeled, and curated the ocular image datasets. The training dataset has been utilized to train the algorithm, with the initial focus on learning normal ocular sonographic anatomy. Our next steps will involve refining the architecture of the developed network to achieve the best fit for the distribution and validation.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Aug 01, 2023
Accession Number
AD1216445

Entities

People

  • Srikar Adhikari

Organizations

  • University of Arizona

Tags

DTIC Thesaurus Topics

  • Algorithms
  • Artificial Intelligence
  • Artificial Intelligence Computing
  • Artificial Intelligence Software
  • Data Analysis
  • Data Curation
  • Data Mining
  • Deep Learning
  • Detection
  • Eye Diseases
  • Health Services
  • Institutional Review Board
  • Machine Learning
  • Medical Personnel
  • Neural Networks
  • Retinal Diseases
  • Training

Fields of Study

  • Medicine

Readers

  • Medical Imaging.
  • Neural Network Machine Learning.
  • Vision Science/Vision Psychology/Cognitive Neuroscience.

Technology Areas

  • AI & ML
  • AI & ML - Neural Networks