Endometriosis detection on ultrasound videos - Sorbonne Université
Conference Poster Year : 2024

Endometriosis detection on ultrasound videos

Abstract

Endometriosis is a chronic condition where tissue similar to the uterine lining grows outside the uterus, affecting up to 10% of reproductive-aged women. It causes severe pain during periods, sexual intercourse, and other activities, and its diagnosis typically requires invasive methods like laparoscopy and biopsy. Our project aims to develop a model using YOLO v8n to detect endometriosis lesions on the rectum via ultrasound, assisting medical professionals in identifying the condition. The model is trained on 17 ultrasound videos, utilizing cross-validation, data augmentation, and key performance metrics such as precision and recall for classification, and mAP for object detection.
Fichier principal
Vignette du fichier
Endometriosis_detection_on_ultrasound_videos.pdf (2.89 Mo) Télécharger le fichier
Origin Files produced by the author(s)

Dates and versions

hal-04706326 , version 1 (23-09-2024)

Identifiers

  • HAL Id : hal-04706326 , version 1

Cite

Olena Verbytska, Baptiste Gregorutti, Martine Valiere. Endometriosis detection on ultrasound videos. Joint DFH/UFA workshop on AI in Medicine: Optimised Trials with Machine Learning, Sep 2024, Paris, France. ⟨hal-04706326⟩
0 View
0 Download

Share

More