DCNN-MF-SP

Evaluation of Deep Convolutional Neural Networks for Automatic Classification of Common Maternal Fetal Ultrasound Planes

---

• 1. Dataset
• 2. Dependent libraries
• 3. Running

---

1. Dataset

Fetal_Planes_DB

2. Dependent libraries

tensorflow >= 2.4
keras >= 2.4
...
...
...
wandb
ipdb
pandas
openpyxl
scipy
pydot
graphviz

If tensorflow cannot use gpu acceleration due to dependencies such as cuda or cudnn, or other dependent library problems, you can use the docker image yulv/tensorflow2 of my configuration environment.

3. Running

CUDA_VISIBLE_DEVICES=0 python main.py --model_type base_model --model_name base_model

CUDA_VISIBLE_DEVICES=1 python main.py --model_type DCNN --model_name VGG19

For more experimental run scripts, see running.sh.