Showing 1–2 of 2 results for author: Lakhani, P
-
Generalization of Artificial Intelligence Models in Medical Imaging: A Case-Based Review
Authors:
Rishi Gadepally,
Andrew Gomella,
Eric Gingold,
Paras Lakhani
Abstract:
The discussions around Artificial Intelligence (AI) and medical imaging are centered around the success of deep learning algorithms. As new algorithms enter the market, it is important for practicing radiologists to understand the pitfalls of various AI algorithms. This entails having a basic understanding of how algorithms are developed, the kind of data they are trained on, and the settings in w…
▽ More
The discussions around Artificial Intelligence (AI) and medical imaging are centered around the success of deep learning algorithms. As new algorithms enter the market, it is important for practicing radiologists to understand the pitfalls of various AI algorithms. This entails having a basic understanding of how algorithms are developed, the kind of data they are trained on, and the settings in which they will be deployed. As with all new technologies, use of AI should be preceded by a fundamental understanding of the risks and benefits to those it is intended to help. This case-based review is intended to point out specific factors practicing radiologists who intend to use AI should consider.
△ Less
Submitted 15 November, 2022;
originally announced November 2022.
-
Automated Segmentation of Vertebrae on Lateral Chest Radiography Using Deep Learning
Authors:
Sanket Badhe,
Varun Singh,
Joy Li,
Paras Lakhani
Abstract:
The purpose of this study is to develop an automated algorithm for thoracic vertebral segmentation on chest radiography using deep learning. 124 de-identified lateral chest radiographs on unique patients were obtained. Segmentations of visible vertebrae were manually performed by a medical student and verified by a board-certified radiologist. 74 images were used for training, 10 for validation, a…
▽ More
The purpose of this study is to develop an automated algorithm for thoracic vertebral segmentation on chest radiography using deep learning. 124 de-identified lateral chest radiographs on unique patients were obtained. Segmentations of visible vertebrae were manually performed by a medical student and verified by a board-certified radiologist. 74 images were used for training, 10 for validation, and 40 were held out for testing. A U-Net deep convolutional neural network was employed for segmentation, using the sum of dice coefficient and binary cross-entropy as the loss function. On the test set, the algorithm demonstrated an average dice coefficient value of 90.5 and an average intersection-over-union (IoU) of 81.75. Deep learning demonstrates promise in the segmentation of vertebrae on lateral chest radiography.
△ Less
Submitted 5 January, 2020;
originally announced January 2020.
