Multiclass-U-Net-for-Fetal-Feature-Extraction

Built a Multiclass U-Net Model for deep learning based segmentation mode capable of identifying the structures of the fetal transventricular plane Upon using the "ipynb" notebook the results can be following results can be achieved

Multiclass U-Net Model

\

The whole model has been implemented in the ipynb file from scratch with the help of TensorFlow 2.0 and Keras

For judging the performance Dice coefficient was used

The formula for Dice coefficient is = 2 * the Area of Overlap divided by the total number of pixels in both images. and is used mostly for segmentation problems

Dice Coefficient Graph Achieved

Image - Labels - Prediction