Interview with Krish Suresh, MD, and Matthew G. Crowson, MD, authors of Making Use of Artificial Intelligence–Generated Synthetic Tympanic Membrane Images. Hosted by Paul C. Bryson, MD, MBA.
Related Content:
Making Use of Artificial Intelligence–Generated Synthetic Tympanic Membrane Images
https://jamanetwork.com/journals/jamaotolaryngology/fullarticle/10.1001/jamaoto.2023.0218?utm_campaign=podcasts&utm_medium=social&utm_source=youtube&utm_content=public
Podcast was originally published on March 30, 2023.
https://edhub.ama-assn.org/jn-learning/audio-player/10.1001/jamaoto.2023.424?utm_campaign=podcasts&utm_medium=social&utm_source=youtube&utm_content=public
———————————————————————————
Follow JAMA Otolaryngology Author Interviews on:
Youtube:
Apple Podcasts: https://podcasts.apple.com/us/podcast/jama-otolaryngology-head-neck-surgery-author-interviews/id1227003951
Spotify: https://open.spotify.com/show/6NUICWavX0TK1KSUv4X7e7
Pocket Casts: https://pca.st/cn4ju2x9
Your podcast app of choice: https://pod.link/qE4lq
Website: https://jamanetwork.com/journals/jamaotolaryngology/pages/jama-otolaryngology-author-interviews
———————————————————————————
For more from JAMA
• https://www.jama.com
• https://www.facebook.com/JAMAJournal
• https://www.twitter.com/JAMA_Current
• https://www.linkedin.com/company/jamanetwork
Follow the JAMA Network
• https://www.jamanetwork.com
• http://www.jamanetworkaudio.com
• https://www.facebook.com/JAMANetwork
• https://www.twitter.com/JAMANetwork
• https://www.instagram.com/JAMANetwork
• https://www.pinterest.com/JAMANetwork
source
date: 2024-07-29 19:53:21
duration: 00:18:07
author: UC4p6bEngiRCN7TyTxZVJ3OQ
Making Use of AI-Generated Synthetic Tympanic Membrane Images
In this podcast episode, Dr. Paul Bryson interviews Dr. Matt Crowson and Dr. Krish Suresh, authors of a research letter published in JAMA Otolaryngology–Head & Neck Surgery, about their use of Artificial Intelligence (AI)-generated synthetic tympanic membrane images to train diagnostic models for detecting middle ear infections. The authors demonstrate that these synthetic images can be used to develop a model that accurately diagnoses middle ear infections, outperforming a model trained on real images.
The AI technology, generative adversarial networks (GANs), generates realistic images that look like real tympanic membrane images, but are actually fake. The synthetic images are then used to train a diagnostic model that can identify abnormal or normal tympanic membrane images. The model is tested on a human cohort, and the results show that it can accurately diagnose middle ear infections with high accuracy.
The authors discuss the potential of this technology to be used in other medical specialties, such as cardiology and pulmonology, to develop diagnostic models for other conditions. They also highlight the challenges and limitations of using AI-generated synthetic data, including the need for careful validation and the potential for biased training data to produce biased results.
The podcast concludes with a discussion on the importance of inter-institutional collaboration, open data, and data sharing to advance the field of AI in medicine. The authors also emphasize the potential for non-technical experts to partner with data scientists to develop innovative applications of AI in medicine. Overall, this podcast provides valuable insights into the use of AI-generated synthetic data in medical imaging and highlights the potential for this technology to improve patient care and outcomes.
