EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

A Study on the Effectiveness of Deep Learning-Based Anomaly Detection Methods for Breast Ultrasonography

Changhee Yun, Bomi Eom iD, Sungjun Park, Chanho Kim, Dohwan Kim iD, Farah Jabeen, Won Hwa Kim, Hye Jung Kim, Jaeil Kim iD

Sensors · 2023

DOI: 10.3390/s23052864 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Sensors, 2023. Verfasst von Changhee Yun und 8 weiteren Autor:innen. Beteiligte Einrichtung: National Information Society Agency, Daegu 41068, Republic of Korea. Hauptthemen laut MEDLINE: Ultrasonography, Mammary, Deep Learning. Die Arbeit wurde bisher 7-mal zitiert (Crossref) und nennt 31 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Kyungpook National University Research Fund.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

In the medical field, it is delicate to anticipate good performance in using deep learning due to the lack of large-scale training data and class imbalance. In particular, ultrasound, which is a key breast cancer diagnosis method, is delicate to diagnose accurately as the quality and interpretation of images can vary depending on the operator’s experience and proficiency. Therefore, computer-aided diagnosis technology can facilitate diagnosis by visualizing abnormal information such as tumors and masses in ultrasound images. In this study, we implemented deep learning-based anomaly detection methods for breast ultrasound images and validated their effectiveness in detecting abnormal regions. Herein, we specifically compared the sliced-Wasserstein autoencoder with two representative unsupervised learning models autoencoder and variational autoencoder. The anomalous region detection performance is estimated with the normal region labels. Our experimental results showed that the sliced-Wasserstein autoencoder model outperformed the anomaly detection performance of others. However, anomaly detection using the reconstruction-based approach may not be effective because of the occurrence of numerous false-positive values. In the following studies, reducing these false positives becomes an important challenge.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Ultrasonography, MammaryDeep LearningFemaleHumansImage Processing, Computer-AssistedUltrasonographyDiagnosis, Computer-Assisted

Stichwörter der Autor:innen

anomaly detectionautoencoderbreast cancerdeep learningultrasonography

Autor:innen und Institutionen

  • National Information Society Agency, Daegu 41068, Republic of Korea.
  • School of Computer Science and Engineering, Kyungpook National University, Daegu 41566, Republic of Korea.
  • Department of Artificial Intelligence, Kyungpook National University, Daegu 41566, Republic of Korea.
  • Department of Radiology, Kyungpook National University Chilgok Hospital, Kyungpook National University, Daegu 41404, Republic of Korea.

ORCID-Profile: Bomi Eom, Dohwan Kim, Jaeil Kim

Förderung

  • Kyungpook National University Research Fund

Zitieren

APA

Yun, C., Eom, B., Park, S., Kim, C., Kim, D., Jabeen, F., Kim, W. H., Kim, H. J., & Kim, J. (2023). A Study on the Effectiveness of Deep Learning-Based Anomaly Detection Methods for Breast Ultrasonography. Sensors. https://doi.org/10.3390/s23052864

Vancouver

Yun C, Eom B, Park S, Kim C, Kim D, Jabeen F, et al. A Study on the Effectiveness of Deep Learning-Based Anomaly Detection Methods for Breast Ultrasonography. Sensors. 2023. doi:10.3390/s23052864

BibTeX
@article{Yun2023,
  author = {Yun, Changhee and Eom, Bomi and Park, Sungjun and Kim, Chanho and Kim, Dohwan and Jabeen, Farah and Kim, Won Hwa and Kim, Hye Jung and Kim, Jaeil},
  title = {A Study on the Effectiveness of Deep Learning-Based Anomaly Detection Methods for Breast Ultrasonography},
  journal = {Sensors},
  year = {2023},
  doi = {10.3390/s23052864}
}
Kostenlos starten →