What question did this study set out to answer?

To develop and evaluate a deep learning computer-aided diagnosis system for classifying thyroid nodules as benign or malignant on ultrasound imaging.

July 10, 2017Journal of Digital ImagingOpen Access

Thyroid Nodule Classification in Ultrasound Images by Fine-Tuning Deep Convolutional Neural Network

View Full Paper

Ask AI

Bookmark

Authors

JCJianning ChiShenyang Aerospace University EWEkta WaliaUniversity of Saskatchewan PBPaul BabynUniversity of Saskatchewan

Discussion

Member takes

Implication

Model evaluation study demonstrates high diagnostic accuracy of deep neural networks in ultrasound thyroid nodule classification, indicating potential to reduce unnecessary invasive biopsies.

Key Points

To develop and evaluate a deep learning computer-aided diagnosis system for classifying thyroid nodules as benign or malignant on ultrasound imaging.
Calibrated image scales and filtered artifacts during ultrasound pre-processing.
Fine-tuned a pre-trained GoogLeNet architecture for high-level image feature extraction.
Classified extracted features using a Cost-sensitive Random Forest across both an open-access database and a local health region database.
Attained 98.29% classification accuracy, 99.10% sensitivity, and 93.90% specificity on the open-access database (Pedraza et al.).
Attained 96.34% classification accuracy, 86.00% sensitivity, and 99.00% specificity on the local health region database.

Cite This Study

Chi et al. (2017) studied this question.

synapsesocial.com/papers/6a19df41a2165c1276df25db https://doi.org/10.1007/s10278-017-9997-y

View Full Paper

Ask AI

Bookmark

Implication

Model evaluation study demonstrates high diagnostic accuracy of deep neural networks in ultrasound thyroid nodule classification, indicating potential to reduce unnecessary invasive biopsies.

Thyroid Nodule Classification in Ultrasound Images by Fine-Tuning Deep Convolutional Neural Network

Discussion

Member takes

Implication

Key Points

Cite This Study

Thyroid Nodule Classification in Ultrasound Images by Fine-Tuning Deep Convolutional Neural Network

Discussion

Member takes

Implication

Key Points

Cite This Study

Also Consider

Also Consider