1. Detection of Ovarian Cyst in Ultrasound Images Using Fine-Tuned VGG-16 Deep Learning Network
- Author
-
Vaishali Chaudhry, Sakshi Srivastava, Prince Kumar, and Anuj Singh
- Subjects
Infertility ,medicine.medical_specialty ,Ovarian cyst ,business.industry ,Deep learning ,Ultrasound ,medicine ,Cyst ,Artificial intelligence ,Radiology ,medicine.disease ,business ,Ultrasound test - Abstract
Ovaries play a vital role in the female reproductive system as they are responsible for the production of egg or ovum required during the fertilization. The female ovaries very often get affected with cyst. An enlarged ovarian cyst can lead to torsion, infertility and even cancer. Therefore, it is very important to diagnose it as soon as possible. For the diagnosis of an ovarian cyst, ultrasound test is conducted. We collected the sample ultrasound images of ovaries of different women and detected whether ovarian cyst is present or not. The proposed work employs the traditional VGG-16 model fine-tuned with our very own dataset of ultrasound images. A VGG-16 model is a 16-layer deep learning neural network trained on ImageNet dataset. Fine-tuning is done by modifying the last four layers of VGG-16 network. Our model is able to determine whether the ultrasound images shows ovarian cyst or not. An accuracy of 92.11% is obtained. The accuracy and loss curves are also plotted for the proposed model.
- Published
- 2020