Search Articles

View query in Help articles search

Search Results (1 to 9 of 9 Results)

Download search results: CSV END BibTex RIS


Discrimination of Radiologists' Experience Level Using Eye-Tracking Technology and Machine Learning: Case Study

Discrimination of Radiologists' Experience Level Using Eye-Tracking Technology and Machine Learning: Case Study

We curate the traditional feature sets to those that could be practically acquired without laborious manual ground truthing of exams, as this would permit large-scale deployment of this technology to health care institutions. To highlight the use of eye-tracking data and artificial intelligence, we term our general approach “biometric radiology artificial intelligence.”

Stanford Martinez, Carolina Ramirez-Tamayo, Syed Hasib Akhter Faruqui, Kal Clark, Adel Alaeddini, Nicholas Czarnek, Aarushi Aggarwal, Sahra Emamzadeh, Jeffrey R Mock, Edward J Golob

JMIR Form Res 2025;9:e53928

Impact of Gold-Standard Label Errors on Evaluating Performance of Deep Learning Models in Diabetic Retinopathy Screening: Nationwide Real-World Validation Study

Impact of Gold-Standard Label Errors on Evaluating Performance of Deep Learning Models in Diabetic Retinopathy Screening: Nationwide Real-World Validation Study

A better understanding of the reasons for the performance gap so as to enhance the real-world performance of DL-based models could help promote the implementation of AI on the ground, but related evidence is scarce. In the development of medical AI, human expert labels usually serve as the gold standard to compare with and evaluate model results [11]. The accuracy of AI model–based tasks critically depends on high-quality labels.

Yueye Wang, Xiaotong Han, Cong Li, Lixia Luo, Qiuxia Yin, Jian Zhang, Guankai Peng, Danli Shi, Mingguang He

J Med Internet Res 2024;26:e52506

Developing the DIGIFOOD Dashboard to Monitor the Digitalization of Local Food Environments: Interdisciplinary Approach

Developing the DIGIFOOD Dashboard to Monitor the Digitalization of Local Food Environments: Interdisciplinary Approach

Moreover, several public health studies have found Google data to be a reliable source for identifying food retailers [27-30] comparable to in-person ground-truthing audits. Thus, it is likely that most food outlets have a Google Maps business listing and can be representative of the local food environment. Relevant search terms (n=120) representative of the food environment were obtained from a list of over 4000 Google My Business categories, publicly available online [31].

Si Si Jia, Xinwei Luo, Alice Anne Gibson, Stephanie Ruth Partridge

JMIR Public Health Surveill 2024;10:e59924

Artificial Intelligence Algorithms in Health Care: Is the Current Food and Drug Administration Regulation Sufficient?

Artificial Intelligence Algorithms in Health Care: Is the Current Food and Drug Administration Regulation Sufficient?

During the local and national testing and retraining, ground-truthing, or annotation of data to compare with algorithm predictions, is an essential process during model evolution. While fully automated methods exist, typically achievable in binary classification tasks with well-structured data, more complex tasks such as segmentation tasks (eg, identification of a lesion on a scan) will require manual labeling by a specialist.

Meghavi Mashar, Shreya Chawla, Fangyue Chen, Baker Lubwama, Kyle Patel, Mihir A Kelshiker, Patrik Bachtiger, Nicholas S Peters

JMIR AI 2023;2:e42940

Acceptability of an In-home Multimodal Sensor Platform for Parkinson Disease: Nonrandomized Qualitative Study

Acceptability of an In-home Multimodal Sensor Platform for Parkinson Disease: Nonrandomized Qualitative Study

Movement Indoor localization Movement Indoor localization Silhouette outline 2 D and 3 D bounding boxes around participant Humidity Temperature Light Pressure Detecting movement Use of kettle, toaster, television, washing machine, refrigerator, and microwave Use of mains electricity Use of toilet and taps in bathrooms High-resolution video data can be used to add additional objective evidence about a symptom at the time when another passive sensor is collecting data, evidence that is called ground truth.

Catherine Morgan, Emma L Tonkin, Ian Craddock, Alan L Whone

JMIR Hum Factors 2022;9(3):e36370

Evaluating Closures of Fresh Fruit and Vegetable Vendors During the COVID-19 Pandemic: Methodology and Preliminary Results Using Omnidirectional Street View Imagery

Evaluating Closures of Fresh Fruit and Vegetable Vendors During the COVID-19 Pandemic: Methodology and Preliminary Results Using Omnidirectional Street View Imagery

Similar to Google Street View, Apple Look Around relies upon ground surveys conducted by commissioned vehicles to collect geographic imagery data [23]. Select areas of Park Slope without 2019 time-stamped Google Street View data were supplemented with 2019 time-stamped Apple Look Around imagery to complete the data collection for this neighborhood. Following the 2019 Google Street View data extraction, an in-person check was conducted for each fresh fruit and vegetable vendor between June and July of 2020.

Shahmir H H. Ali, Valerie M Imbruce, Rienna G Russo, Samuel Kaplan, Kaye Stevenson, Tamar Adjoian Mezzacca, Victoria Foster, Ashley Radee, Stella Chong, Felice Tsui, Julie Kranick, Stella S Yi

JMIR Form Res 2021;5(2):e23870

Locating Medical and Recreational Cannabis Outlets for Research Purposes: Online Methods and Observational Study

Locating Medical and Recreational Cannabis Outlets for Research Purposes: Online Methods and Observational Study

It should also be noted that observational methods alone, such as ground truthing, where observers would drive every street in an area to locate targeted retailers (eg, locating vape shops [18]), would be unfeasible, given there are 4751 square miles in Los Angeles (85% of which are land), and also that many cannabis outlets identified from the online sources were unrecognizable during observations as outlets that sold cannabis (ie, 27.7% had no signage indicating the outlet sold cannabis).

Eric R R Pedersen, Caislin Firth, Jennifer Parker, Regina A Shih, Steven Davenport, Anthony Rodriguez, Michael S Dunbar, Lisa Kraus, Joan S Tucker, Elizabeth J D'Amico

J Med Internet Res 2020;22(2):e16853

Establishing a Demographic, Development and Environmental Geospatial Surveillance Platform in India: Planning and Implementation

Establishing a Demographic, Development and Environmental Geospatial Surveillance Platform in India: Planning and Implementation

The geospatial data so constructed was subsequently reassessed for land parcel position (location, size, and shape) and attribute accuracy through ground-truthing–based verification exercises. Temporal changes that emerged during the course of the data construction were also incorporated in this exercise. The methodology for verification of geospatial datasets through community-based ground-truthing was formulated through a pilot study conducted at 6 surveillance villages.

Shikha Dixit, Narendra K Arora, Atiqur Rahman, Natasha J Howard, Rakesh K Singh, Mayur Vaswani, Manoja K Das, Faruqueuddin Ahmed, Prashant Mathur, Nikhil Tandon, Rajib Dasgupta, Sanjay Chaturvedi, Jaishri Jethwaney, Suresh Dalpath, Rajendra Prashad, Rakesh Kumar, Rakesh Gupta, Laurette Dube, Mark Daniel

JMIR Public Health Surveill 2018;4(4):e66